PLAY PODCASTS
Daily Paper Cast

Daily Paper Cast

2,516 episodes — Page 49 of 51

Ep 116SAMURAI: Adapting Segment Anything Model for Zero-Shot Visual Tracking with Motion-Aware Memory

🤗 Paper Upvotes: 12 | cs.CV Authors: Cheng-Yen Yang, Hsiang-Wei Huang, Wenhao Chai, Zhongyu Jiang, Jenq-Neng Hwang Title: SAMURAI: Adapting Segment Anything Model for Zero-Shot Visual Tracking with Motion-Aware Memory Arxiv: http://arxiv.org/abs/2411.11922v1 Abstract: The Segment Anything Model 2 (SAM 2) has demonstrated strong performance in object segmentation tasks but faces challenges in visual object tracking, particularly when managing crowded scenes with fast-moving or self-occluding objects. Furthermore, the fixed-window memory approach in the original model does not consider the quality of memories selected to condition the image features for the next frame, leading to error propagation in videos. This paper introduces SAMURAI, an enhanced adaptation of SAM 2 specifically designed for visual object tracking. By incorporating temporal motion cues with the proposed motion-aware memory selection mechanism, SAMURAI effectively predicts object motion and refines mask selection, achieving robust, accurate tracking without the need for retraining or fine-tuning. SAMURAI operates in real-time and demonstrates strong zero-shot performance across diverse benchmark datasets, showcasing its ability to generalize without fine-tuning. In evaluations, SAMURAI achieves significant improvements in success rate and precision over existing trackers, with a 7.1% AUC gain on LaSOT$_{\text{ext}}$ and a 3.5% AO gain on GOT-10k. Moreover, it achieves competitive results compared to fully supervised methods on LaSOT, underscoring its robustness in complex tracking scenarios and its potential for real-world applications in dynamic environments. Code and results are available at https://github.com/yangchris11/samurai.

Nov 22, 202422 min

Ep 115Is Your LLM Secretly a World Model of the Internet? Model-Based Planning for Web Agents

🤗 Paper Upvotes: 9 | cs.AI Authors: Yu Gu, Boyuan Zheng, Boyu Gou, Kai Zhang, Cheng Chang, Sanjari Srivastava, Yanan Xie, Peng Qi, Huan Sun, Yu Su Title: Is Your LLM Secretly a World Model of the Internet? Model-Based Planning for Web Agents Arxiv: http://arxiv.org/abs/2411.06559v1 Abstract: Language agents have demonstrated promising capabilities in automating web-based tasks, though their current reactive approaches still underperform largely compared to humans. While incorporating advanced planning algorithms, particularly tree search methods, could enhance these agents' performance, implementing tree search directly on live websites poses significant safety risks and practical constraints due to irreversible actions such as confirming a purchase. In this paper, we introduce a novel paradigm that augments language agents with model-based planning, pioneering the innovative use of large language models (LLMs) as world models in complex web environments. Our method, WebDreamer, builds on the key insight that LLMs inherently encode comprehensive knowledge about website structures and functionalities. Specifically, WebDreamer uses LLMs to simulate outcomes for each candidate action (e.g., "what would happen if I click this button?") using natural language descriptions, and then evaluates these imagined outcomes to determine the optimal action at each step. Empirical results on two representative web agent benchmarks with online interaction -- VisualWebArena and Mind2Web-live -- demonstrate that WebDreamer achieves substantial improvements over reactive baselines. By establishing the viability of LLMs as world models in web environments, this work lays the groundwork for a paradigm shift in automated web interaction. More broadly, our findings open exciting new avenues for future research into 1) optimizing LLMs specifically for world modeling in complex, dynamic environments, and 2) model-based speculative planning for language agents.

Nov 22, 202421 min

Ep 114When Precision Meets Position: BFloat16 Breaks Down RoPE in Long-Context Training

🤗 Paper Upvotes: 7 | cs.CL Authors: Haonan Wang, Qian Liu, Chao Du, Tongyao Zhu, Cunxiao Du, Kenji Kawaguchi, Tianyu Pang Title: When Precision Meets Position: BFloat16 Breaks Down RoPE in Long-Context Training Arxiv: http://arxiv.org/abs/2411.13476v1 Abstract: Extending context window sizes allows large language models (LLMs) to process longer sequences and handle more complex tasks. Rotary Positional Embedding (RoPE) has become the de facto standard due to its relative positional encoding properties that benefit long-context training. However, we observe that using RoPE with BFloat16 format results in numerical issues, causing it to deviate from its intended relative positional encoding, especially in long-context scenarios. This issue arises from BFloat16's limited precision and accumulates as context length increases, with the first token contributing significantly to this problem. To address this, we develop AnchorAttention, a plug-and-play attention method that alleviates numerical issues caused by BFloat16, improves long-context capabilities, and speeds up training. AnchorAttention reduces unnecessary attention computations, maintains semantic coherence, and boosts computational efficiency by treating the first token as a shared anchor with a consistent position ID, making it visible to all documents within the training context. Experiments on three types of LLMs demonstrate that AnchorAttention significantly improves long-context performance and reduces training time by over 50\% compared to standard full attention mechanisms, while preserving the original LLM's capabilities on general tasks. Our code is available at https://github.com/haonan3/AnchorContext.

Nov 22, 202424 min

Ep 113Stylecodes: Encoding Stylistic Information For Image Generation

🤗 Paper Upvotes: 6 | cs.CV Authors: Ciara Rowles Title: Stylecodes: Encoding Stylistic Information For Image Generation Arxiv: http://arxiv.org/abs/2411.12811v1 Abstract: Diffusion models excel in image generation, but controlling them remains a challenge. We focus on the problem of style-conditioned image generation. Although example images work, they are cumbersome: srefs (style-reference codes) from MidJourney solve this issue by expressing a specific image style in a short numeric code. These have seen widespread adoption throughout social media due to both their ease of sharing and the fact they allow using an image for style control, without having to post the source images themselves. However, users are not able to generate srefs from their own images, nor is the underlying training procedure public. We propose StyleCodes: an open-source and open-research style encoder architecture and training procedure to express image style as a 20-symbol base64 code. Our experiments show that our encoding results in minimal loss in quality compared to traditional image-to-style techniques.

Nov 22, 202421 min

Ep 112ViBe: A Text-to-Video Benchmark for Evaluating Hallucination in Large Multimodal Models

🤗 Paper Upvotes: 3 | cs.CV, cs.AI Authors: Vipula Rawte, Sarthak Jain, Aarush Sinha, Garv Kaushik, Aman Bansal, Prathiksha Rumale Vishwanath, Samyak Rajesh Jain, Aishwarya Naresh Reganti, Vinija Jain, Aman Chadha, Amit P. Sheth, Amitava Das Title: ViBe: A Text-to-Video Benchmark for Evaluating Hallucination in Large Multimodal Models Arxiv: http://arxiv.org/abs/2411.10867v1 Abstract: Latest developments in Large Multimodal Models (LMMs) have broadened their capabilities to include video understanding. Specifically, Text-to-video (T2V) models have made significant progress in quality, comprehension, and duration, excelling at creating videos from simple textual prompts. Yet, they still frequently produce hallucinated content that clearly signals the video is AI-generated. We introduce ViBe: a large-scale Text-to-Video Benchmark of hallucinated videos from T2V models. We identify five major types of hallucination: Vanishing Subject, Numeric Variability, Temporal Dysmorphia, Omission Error, and Physical Incongruity. Using 10 open-source T2V models, we developed the first large-scale dataset of hallucinated videos, comprising 3,782 videos annotated by humans into these five categories. ViBe offers a unique resource for evaluating the reliability of T2V models and provides a foundation for improving hallucination detection and mitigation in video generation. We establish classification as a baseline and present various ensemble classifier configurations, with the TimeSFormer + CNN combination yielding the best performance, achieving 0.345 accuracy and 0.342 F1 score. This benchmark aims to drive the development of robust T2V models that produce videos more accurately aligned with input prompts.

Nov 22, 202423 min

Ep 111Loss-to-Loss Prediction: Scaling Laws for All Datasets

🤗 Paper Upvotes: 2 | cs.LG, cs.AI, cs.CL, stat.ML Authors: David Brandfonbrener, Nikhil Anand, Nikhil Vyas, Eran Malach, Sham Kakade Title: Loss-to-Loss Prediction: Scaling Laws for All Datasets Arxiv: http://arxiv.org/abs/2411.12925v1 Abstract: While scaling laws provide a reliable methodology for predicting train loss across compute scales for a single data distribution, less is known about how these predictions should change as we change the distribution. In this paper, we derive a strategy for predicting one loss from another and apply it to predict across different pre-training datasets and from pre-training data to downstream task data. Our predictions extrapolate well even at 20x the largest FLOP budget used to fit the curves. More precisely, we find that there are simple shifted power law relationships between (1) the train losses of two models trained on two separate datasets when the models are paired by training compute (train-to-train), (2) the train loss and the test loss on any downstream distribution for a single model (train-to-test), and (3) the test losses of two models trained on two separate train datasets (test-to-test). The results hold up for pre-training datasets that differ substantially (some are entirely code and others have no code at all) and across a variety of downstream tasks. Finally, we find that in some settings these shifted power law relationships can yield more accurate predictions than extrapolating single-dataset scaling laws.

Nov 22, 202421 min

Ep 110ORID: Organ-Regional Information Driven Framework for Radiology Report Generation

🤗 Paper Upvotes: 2 | cs.CV Authors: Tiancheng Gu, Kaicheng Yang, Xiang An, Ziyong Feng, Dongnan Liu, Weidong Cai Title: ORID: Organ-Regional Information Driven Framework for Radiology Report Generation Arxiv: http://arxiv.org/abs/2411.13025v1 Abstract: The objective of Radiology Report Generation (RRG) is to automatically generate coherent textual analyses of diseases based on radiological images, thereby alleviating the workload of radiologists. Current AI-based methods for RRG primarily focus on modifications to the encoder-decoder model architecture. To advance these approaches, this paper introduces an Organ-Regional Information Driven (ORID) framework which can effectively integrate multi-modal information and reduce the influence of noise from unrelated organs. Specifically, based on the LLaVA-Med, we first construct an RRG-related instruction dataset to improve organ-regional diagnosis description ability and get the LLaVA-Med-RRG. After that, we propose an organ-based cross-modal fusion module to effectively combine the information from the organ-regional diagnosis description and radiology image. To further reduce the influence of noise from unrelated organs on the radiology report generation, we introduce an organ importance coefficient analysis module, which leverages Graph Neural Network (GNN) to examine the interconnections of the cross-modal information of each organ region. Extensive experiments an1d comparisons with state-of-the-art methods across various evaluation metrics demonstrate the superior performance of our proposed method.

Nov 22, 202420 min

Ep 109SymDPO: Boosting In-Context Learning of Large Multimodal Models with Symbol Demonstration Direct Preference Optimization

🤗 Paper Upvotes: 13 | cs.CV Authors: Hongrui Jia, Chaoya Jiang, Haiyang Xu, Wei Ye, Mengfan Dong, Ming Yan, Ji Zhang, Fei Huang, Shikun Zhang Title: SymDPO: Boosting In-Context Learning of Large Multimodal Models with Symbol Demonstration Direct Preference Optimization Arxiv: http://arxiv.org/abs/2411.11909v1 Abstract: As language models continue to scale, Large Language Models (LLMs) have exhibited emerging capabilities in In-Context Learning (ICL), enabling them to solve language tasks by prefixing a few in-context demonstrations (ICDs) as context. Inspired by these advancements, researchers have extended these techniques to develop Large Multimodal Models (LMMs) with ICL capabilities. However, existing LMMs face a critical issue: they often fail to effectively leverage the visual context in multimodal demonstrations and instead simply follow textual patterns. This indicates that LMMs do not achieve effective alignment between multimodal demonstrations and model outputs. To address this problem, we propose Symbol Demonstration Direct Preference Optimization (SymDPO). Specifically, SymDPO aims to break the traditional paradigm of constructing multimodal demonstrations by using random symbols to replace text answers within instances. This forces the model to carefully understand the demonstration images and establish a relationship between the images and the symbols to answer questions correctly. We validate the effectiveness of this method on multiple benchmarks, demonstrating that with SymDPO, LMMs can more effectively understand the multimodal context within examples and utilize this knowledge to answer questions better.

Nov 21, 202425 min

Ep 108Continuous Speculative Decoding for Autoregressive Image Generation

🤗 Paper Upvotes: 13 | cs.CV Authors: Zili Wang, Robert Zhang, Kun Ding, Qi Yang, Fei Li, Shiming Xiang Title: Continuous Speculative Decoding for Autoregressive Image Generation Arxiv: http://arxiv.org/abs/2411.11925v1 Abstract: Continuous-valued Autoregressive (AR) image generation models have demonstrated notable superiority over their discrete-token counterparts, showcasing considerable reconstruction quality and higher generation fidelity. However, the computational demands of the autoregressive framework result in significant inference overhead. While speculative decoding has proven effective in accelerating Large Language Models (LLMs), their adaptation to continuous-valued visual autoregressive models remains unexplored. This work generalizes the speculative decoding algorithm from discrete tokens to continuous space. By analyzing the intrinsic properties of output distribution, we establish a tailored acceptance criterion for the diffusion distributions prevalent in such models. To overcome the inconsistency that occurred in speculative decoding output distributions, we introduce denoising trajectory alignment and token pre-filling methods. Additionally, we identify the hard-to-sample distribution in the rejection phase. To mitigate this issue, we propose a meticulous acceptance-rejection sampling method with a proper upper bound, thereby circumventing complex integration. Experimental results show that our continuous speculative decoding achieves a remarkable $2.33\times$ speed-up on off-the-shelf models while maintaining the output distribution. Codes will be available at https://github.com/MarkXCloud/CSpD

Nov 21, 202422 min

Ep 107ITACLIP: Boosting Training-Free Semantic Segmentation with Image, Text, and Architectural Enhancements

🤗 Paper Upvotes: 11 | cs.CV Authors: M. Arda Aydın, Efe Mert Çırpar, Elvin Abdinli, Gozde Unal, Yusuf H. Sahin Title: ITACLIP: Boosting Training-Free Semantic Segmentation with Image, Text, and Architectural Enhancements Arxiv: http://arxiv.org/abs/2411.12044v1 Abstract: Recent advances in foundational Vision Language Models (VLMs) have reshaped the evaluation paradigm in computer vision tasks. These foundational models, especially CLIP, have accelerated research in open-vocabulary computer vision tasks, including Open-Vocabulary Semantic Segmentation (OVSS). Although the initial results are promising, the dense prediction capabilities of VLMs still require further improvement. In this study, we enhance the semantic segmentation performance of CLIP by introducing new modules and modifications: 1) architectural changes in the last layer of ViT and the incorporation of attention maps from the middle layers with the last layer, 2) Image Engineering: applying data augmentations to enrich input image representations, and 3) using Large Language Models (LLMs) to generate definitions and synonyms for each class name to leverage CLIP's open-vocabulary capabilities. Our training-free method, ITACLIP, outperforms current state-of-the-art approaches on segmentation benchmarks such as COCO-Stuff, COCO-Object, Pascal Context, and Pascal VOC. Our code is available at https://github.com/m-arda-aydn/ITACLIP.

Nov 21, 202419 min

Ep 106FlipSketch: Flipping Static Drawings to Text-Guided Sketch Animations

🤗 Paper Upvotes: 10 | cs.GR, cs.CV Authors: Hmrishav Bandyopadhyay, Yi-Zhe Song Title: FlipSketch: Flipping Static Drawings to Text-Guided Sketch Animations Arxiv: http://arxiv.org/abs/2411.10818v1 Abstract: Sketch animations offer a powerful medium for visual storytelling, from simple flip-book doodles to professional studio productions. While traditional animation requires teams of skilled artists to draw key frames and in-between frames, existing automation attempts still demand significant artistic effort through precise motion paths or keyframe specification. We present FlipSketch, a system that brings back the magic of flip-book animation -- just draw your idea and describe how you want it to move! Our approach harnesses motion priors from text-to-video diffusion models, adapting them to generate sketch animations through three key innovations: (i) fine-tuning for sketch-style frame generation, (ii) a reference frame mechanism that preserves visual integrity of input sketch through noise refinement, and (iii) a dual-attention composition that enables fluid motion without losing visual consistency. Unlike constrained vector animations, our raster frames support dynamic sketch transformations, capturing the expressive freedom of traditional animation. The result is an intuitive system that makes sketch animation as simple as doodling and describing, while maintaining the artistic essence of hand-drawn animation.

Nov 21, 202425 min

Ep 105Soft Robotic Dynamic In-Hand Pen Spinning

🤗 Paper Upvotes: 8 | cs.RO Authors: Yunchao Yao, Uksang Yoo, Jean Oh, Christopher G. Atkeson, Jeffrey Ichnowski Title: Soft Robotic Dynamic In-Hand Pen Spinning Arxiv: http://arxiv.org/abs/2411.12734v1 Abstract: Dynamic in-hand manipulation remains a challenging task for soft robotic systems that have demonstrated advantages in safe compliant interactions but struggle with high-speed dynamic tasks. In this work, we present SWIFT, a system for learning dynamic tasks using a soft and compliant robotic hand. Unlike previous works that rely on simulation, quasi-static actions and precise object models, the proposed system learns to spin a pen through trial-and-error using only real-world data without requiring explicit prior knowledge of the pen's physical attributes. With self-labeled trials sampled from the real world, the system discovers the set of pen grasping and spinning primitive parameters that enables a soft hand to spin a pen robustly and reliably. After 130 sampled actions per object, SWIFT achieves 100% success rate across three pens with different weights and weight distributions, demonstrating the system's generalizability and robustness to changes in object properties. The results highlight the potential for soft robotic end-effectors to perform dynamic tasks including rapid in-hand manipulation. We also demonstrate that SWIFT generalizes to spinning items with different shapes and weights such as a brush and a screwdriver which we spin with 10/10 and 5/10 success rates respectively. Videos, data, and code are available at https://soft-spin.github.io.

Nov 21, 202422 min

Ep 104Building Trust: Foundations of Security, Safety and Transparency in AI

🤗 Paper Upvotes: 8 | cs.CY, cs.AI, cs.CL Authors: Huzaifa Sidhpurwala, Garth Mollett, Emily Fox, Mark Bestavros, Huamin Chen Title: Building Trust: Foundations of Security, Safety and Transparency in AI Arxiv: http://arxiv.org/abs/2411.12275v1 Abstract: This paper explores the rapidly evolving ecosystem of publicly available AI models, and their potential implications on the security and safety landscape. As AI models become increasingly prevalent, understanding their potential risks and vulnerabilities is crucial. We review the current security and safety scenarios while highlighting challenges such as tracking issues, remediation, and the apparent absence of AI model lifecycle and ownership processes. Comprehensive strategies to enhance security and safety for both model developers and end-users are proposed. This paper aims to provide some of the foundational pieces for more standardized security, safety, and transparency in the development and operation of AI models and the larger open ecosystems and communities forming around them.

Nov 21, 202422 min

Ep 103SEAGULL: No-reference Image Quality Assessment for Regions of Interest via Vision-Language Instruction Tuning

🤗 Paper Upvotes: 5 | cs.CV Authors: Zewen Chen, Juan Wang, Wen Wang, Sunhan Xu, Hang Xiong, Yun Zeng, Jian Guo, Shuxun Wang, Chunfeng Yuan, Bing Li, Weiming Hu Title: SEAGULL: No-reference Image Quality Assessment for Regions of Interest via Vision-Language Instruction Tuning Arxiv: http://arxiv.org/abs/2411.10161v1 Abstract: Existing Image Quality Assessment (IQA) methods achieve remarkable success in analyzing quality for overall image, but few works explore quality analysis for Regions of Interest (ROIs). The quality analysis of ROIs can provide fine-grained guidance for image quality improvement and is crucial for scenarios focusing on region-level quality. This paper proposes a novel network, SEAGULL, which can SEe and Assess ROIs quality with GUidance from a Large vision-Language model. SEAGULL incorporates a vision-language model (VLM), masks generated by Segment Anything Model (SAM) to specify ROIs, and a meticulously designed Mask-based Feature Extractor (MFE) to extract global and local tokens for specified ROIs, enabling accurate fine-grained IQA for ROIs. Moreover, this paper constructs two ROI-based IQA datasets, SEAGULL-100w and SEAGULL-3k, for training and evaluating ROI-based IQA. SEAGULL-100w comprises about 100w synthetic distortion images with 33 million ROIs for pre-training to improve the model's ability of regional quality perception, and SEAGULL-3k contains about 3k authentic distortion ROIs to enhance the model's ability to perceive real world distortions. After pre-training on SEAGULL-100w and fine-tuning on SEAGULL-3k, SEAGULL shows remarkable performance on fine-grained ROI quality assessment. Code and datasets are publicly available at the https://github.com/chencn2020/Seagull.

Nov 21, 202420 min

Ep 102Evaluating Tokenizer Performance of Large Language Models Across Official Indian Languages

🤗 Paper Upvotes: 3 | cs.CL, cs.AI Authors: S. Tamang, D. J. Bora Title: Evaluating Tokenizer Performance of Large Language Models Across Official Indian Languages Arxiv: http://arxiv.org/abs/2411.12240v1 Abstract: Large Language Models (LLMs) based on transformer architectures have revolutionized a variety of domains, with tokenization playing a pivotal role in their pre-processing and fine-tuning stages. In multilingual models, particularly those tailored for Indic languages, effective tokenization is crucial for optimizing performance. This paper presents a comprehensive evaluation of tokenizers used by 12 LLMs across all 22 official languages of India, with a focus on comparing the efficiency of their tokenization processes. We employed the Normalized Sequence Length (NSL) as a key metric in our analysis. Our findings reveal that the SUTRA tokenizer outperforms all other models, including several Indic-specific models, excelling in 14 languages. Notable insights include the SUTRA tokenizer's superior handling of Indic languages, GPT-4o's advancement over its predecessor GPT-4 in processing Indian languages, and the limited performance of Project Indus in certain languages. This study underscores the critical importance of developing targeted tokenization strategies for multilingual and Indic-centric models, laying the groundwork for future improvements in tokenizer design to enhance linguistic coverage and model efficiency.

Nov 21, 202424 min

Ep 101Generative World Explorer

🤗 Paper Upvotes: 38 | cs.CV Authors: Taiming Lu, Tianmin Shu, Alan Yuille, Daniel Khashabi, Jieneng Chen Title: Generative World Explorer Arxiv: http://arxiv.org/abs/2411.11844v2 Abstract: Planning with partial observation is a central challenge in embodied AI. A majority of prior works have tackled this challenge by developing agents that physically explore their environment to update their beliefs about the world state. In contrast, humans can $\textit{imagine}$ unseen parts of the world through a mental exploration and $\textit{revise}$ their beliefs with imagined observations. Such updated beliefs can allow them to make more informed decisions, without necessitating the physical exploration of the world at all times. To achieve this human-like ability, we introduce the $\textit{Generative World Explorer (Genex)}$, an egocentric world exploration framework that allows an agent to mentally explore a large-scale 3D world (e.g., urban scenes) and acquire imagined observations to update its belief. This updated belief will then help the agent to make a more informed decision at the current step. To train $\textit{Genex}$, we create a synthetic urban scene dataset, Genex-DB. Our experimental results demonstrate that (1) $\textit{Genex}$ can generate high-quality and consistent observations during long-horizon exploration of a large virtual physical world and (2) the beliefs updated with the generated observations can inform an existing decision-making model (e.g., an LLM agent) to make better plans.

Nov 20, 202421 min

Ep 100BlueLM-V-3B: Algorithm and System Co-Design for Multimodal Large Language Models on Mobile Devices

🤗 Paper Upvotes: 31 | cs.CV, cs.CL Authors: Xudong Lu, Yinghao Chen, Cheng Chen, Hui Tan, Boheng Chen, Yina Xie, Rui Hu, Guanxin Tan, Renshou Wu, Yan Hu, Yi Zeng, Lei Wu, Liuyang Bian, Zhaoxiong Wang, Long Liu, Yanzhou Yang, Han Xiao, Aojun Zhou, Yafei Wen, Xiaoxin Chen, Shuai Ren, Hongsheng Li Title: BlueLM-V-3B: Algorithm and System Co-Design for Multimodal Large Language Models on Mobile Devices Arxiv: http://arxiv.org/abs/2411.10640v1 Abstract: The emergence and growing popularity of multimodal large language models (MLLMs) have significant potential to enhance various aspects of daily life, from improving communication to facilitating learning and problem-solving. Mobile phones, as essential daily companions, represent the most effective and accessible deployment platform for MLLMs, enabling seamless integration into everyday tasks. However, deploying MLLMs on mobile phones presents challenges due to limitations in memory size and computational capability, making it difficult to achieve smooth and real-time processing without extensive optimization. In this paper, we present BlueLM-V-3B, an algorithm and system co-design approach specifically tailored for the efficient deployment of MLLMs on mobile platforms. To be specific, we redesign the dynamic resolution scheme adopted by mainstream MLLMs and implement system optimization for hardware-aware deployment to optimize model inference on mobile phones. BlueLM-V-3B boasts the following key highlights: (1) Small Size: BlueLM-V-3B features a language model with 2.7B parameters and a vision encoder with 400M parameters. (2) Fast Speed: BlueLM-V-3B achieves a generation speed of 24.4 token/s on the MediaTek Dimensity 9300 processor with 4-bit LLM weight quantization. (3) Strong Performance: BlueLM-V-3B has attained the highest average score of 66.1 on the OpenCompass benchmark among models with $\leq$ 4B parameters and surpassed a series of models with much larger parameter sizes (e.g., MiniCPM-V-2.6, InternVL2-8B).

Nov 20, 202419 min

Ep 99Search, Verify and Feedback: Towards Next Generation Post-training Paradigm of Foundation Models via Verifier Engineering

🤗 Paper Upvotes: 13 | cs.AI, cs.CL, stat.ML Authors: Xinyan Guan, Yanjiang Liu, Xinyu Lu, Boxi Cao, Ben He, Xianpei Han, Le Sun, Jie Lou, Bowen Yu, Yaojie Lu, Hongyu Lin Title: Search, Verify and Feedback: Towards Next Generation Post-training Paradigm of Foundation Models via Verifier Engineering Arxiv: http://arxiv.org/abs/2411.11504v1 Abstract: The evolution of machine learning has increasingly prioritized the development of powerful models and more scalable supervision signals. However, the emergence of foundation models presents significant challenges in providing effective supervision signals necessary for further enhancing their capabilities. Consequently, there is an urgent need to explore novel supervision signals and technical approaches. In this paper, we propose verifier engineering, a novel post-training paradigm specifically designed for the era of foundation models. The core of verifier engineering involves leveraging a suite of automated verifiers to perform verification tasks and deliver meaningful feedback to foundation models. We systematically categorize the verifier engineering process into three essential stages: search, verify, and feedback, and provide a comprehensive review of state-of-the-art research developments within each stage. We believe that verifier engineering constitutes a fundamental pathway toward achieving Artificial General Intelligence.

Nov 20, 202421 min

Ep 98AnimateAnything: Consistent and Controllable Animation for Video Generation

🤗 Paper Upvotes: 12 | cs.CV Authors: Guojun Lei, Chi Wang, Hong Li, Rong Zhang, Yikai Wang, Weiwei Xu Title: AnimateAnything: Consistent and Controllable Animation for Video Generation Arxiv: http://arxiv.org/abs/2411.10836v1 Abstract: We present a unified controllable video generation approach AnimateAnything that facilitates precise and consistent video manipulation across various conditions, including camera trajectories, text prompts, and user motion annotations. Specifically, we carefully design a multi-scale control feature fusion network to construct a common motion representation for different conditions. It explicitly converts all control information into frame-by-frame optical flows. Then we incorporate the optical flows as motion priors to guide final video generation. In addition, to reduce the flickering issues caused by large-scale motion, we propose a frequency-based stabilization module. It can enhance temporal coherence by ensuring the video's frequency domain consistency. Experiments demonstrate that our method outperforms the state-of-the-art approaches. For more details and videos, please refer to the webpage: https://yu-shaonian.github.io/Animate_Anything/.

Nov 20, 202422 min

Ep 97Top-$nσ$: Not All Logits Are You Need

🤗 Paper Upvotes: 12 | cs.LG Authors: Chenxia Tang, Jianchun Liu, Hongli Xu, Liusheng Huang Title: Top-$nσ$: Not All Logits Are You Need Arxiv: http://arxiv.org/abs/2411.07641v1 Abstract: Large language models (LLMs) typically employ greedy decoding or low-temperature sampling for reasoning tasks, reflecting a perceived trade-off between diversity and accuracy. We challenge this convention by introducing top-$n\sigma$, a novel sampling method that operates directly on pre-softmax logits by leveraging a statistical threshold. Our key insight is that logits naturally separate into a Gaussian-distributed noisy region and a distinct informative region, enabling efficient token filtering without complex probability manipulations. Unlike existing methods (e.g., top-$p$, min-$p$) that inadvertently include more noise tokens at higher temperatures, top-$n\sigma$ maintains a stable sampling space regardless of temperature scaling. We also provide a theoretical analysis of top-$n\sigma$ to better understand its behavior. The extensive experimental results across four reasoning-focused datasets demonstrate that our method not only outperforms existing sampling approaches but also surpasses greedy decoding, while maintaining consistent performance even at high temperatures.

Nov 20, 202421 min

Ep 96Drowning in Documents: Consequences of Scaling Reranker Inference

🤗 Paper Upvotes: 10 | cs.IR, cs.CL, cs.LG Authors: Mathew Jacob, Erik Lindgren, Matei Zaharia, Michael Carbin, Omar Khattab, Andrew Drozdov Title: Drowning in Documents: Consequences of Scaling Reranker Inference Arxiv: http://arxiv.org/abs/2411.11767v1 Abstract: Rerankers, typically cross-encoders, are often used to re-score the documents retrieved by cheaper initial IR systems. This is because, though expensive, rerankers are assumed to be more effective. We challenge this assumption by measuring reranker performance for full retrieval, not just re-scoring first-stage retrieval. Our experiments reveal a surprising trend: the best existing rerankers provide diminishing returns when scoring progressively more documents and actually degrade quality beyond a certain limit. In fact, in this setting, rerankers can frequently assign high scores to documents with no lexical or semantic overlap with the query. We hope that our findings will spur future research to improve reranking.

Nov 20, 202421 min

Ep 95SlimLM: An Efficient Small Language Model for On-Device Document Assistance

🤗 Paper Upvotes: 10 | cs.CL Authors: Thang M. Pham, Phat T. Nguyen, Seunghyun Yoon, Viet Dac Lai, Franck Dernoncourt, Trung Bui Title: SlimLM: An Efficient Small Language Model for On-Device Document Assistance Arxiv: http://arxiv.org/abs/2411.09944v1 Abstract: While small language models (SLMs) show promises for mobile deployment, their real-world performance and applications on smartphones remains underexplored. We present SlimLM, a series of SLMs optimized for document assistance tasks on mobile devices. Through extensive experiments on a Samsung Galaxy S24, we identify the optimal trade-offs between model size (ranging from 125M to 7B parameters), context length, and inference time for efficient on-device processing. SlimLM is pre-trained on SlimPajama-627B and fine-tuned on DocAssist, our constructed dataset for summarization, question answering and suggestion tasks. Our smallest model demonstrates efficient performance on S24, while larger variants offer enhanced capabilities within mobile constraints. We evaluate SlimLM against existing SLMs, showing comparable or superior performance and offering a benchmark for future research in on-device language models. We also provide an Android application, offering practical insights into SLM deployment. Our findings provide valuable insights and illuminate the capabilities of running advanced language models on high-end smartphones, potentially reducing server costs and enhancing privacy through on-device processing.

Nov 20, 202425 min

Ep 94Awaker2.5-VL: Stably Scaling MLLMs with Parameter-Efficient Mixture of Experts

🤗 Paper Upvotes: 8 | cs.CV Authors: Jinqiang Long, Yanqi Dai, Guoxing Yang, Hongpeng Lin, Nanyi Fei, Yizhao Gao, Zhiwu Lu Title: Awaker2.5-VL: Stably Scaling MLLMs with Parameter-Efficient Mixture of Experts Arxiv: http://arxiv.org/abs/2411.10669v1 Abstract: As the research of Multimodal Large Language Models (MLLMs) becomes popular, an advancing MLLM model is typically required to handle various textual and visual tasks (e.g., VQA, Detection, OCR, and ChartQA) simultaneously for real-world applications. However, due to the significant differences in representation and distribution among data from various tasks, simply mixing data of all tasks together leads to the well-known``multi-task conflict" issue, resulting in performance degradation across various tasks. To address this issue, we propose Awaker2.5-VL, a Mixture of Experts~(MoE) architecture suitable for MLLM, which acquires the multi-task capabilities through multiple sparsely activated experts. To speed up the training and inference of Awaker2.5-VL, each expert in our model is devised as a low-rank adaptation (LoRA) structure. Extensive experiments on multiple latest benchmarks demonstrate the effectiveness of Awaker2.5-VL. The code and model weight are released in our Project Page: https://github.com/MetabrainAGI/Awaker.

Nov 20, 202419 min

Ep 93SmoothCache: A Universal Inference Acceleration Technique for Diffusion Transformers

🤗 Paper Upvotes: 8 | cs.LG Authors: Joseph Liu, Joshua Geddes, Ziyu Guo, Haomiao Jiang, Mahesh Kumar Nandwana Title: SmoothCache: A Universal Inference Acceleration Technique for Diffusion Transformers Arxiv: http://arxiv.org/abs/2411.10510v1 Abstract: Diffusion Transformers (DiT) have emerged as powerful generative models for various tasks, including image, video, and speech synthesis. However, their inference process remains computationally expensive due to the repeated evaluation of resource-intensive attention and feed-forward modules. To address this, we introduce SmoothCache, a model-agnostic inference acceleration technique for DiT architectures. SmoothCache leverages the observed high similarity between layer outputs across adjacent diffusion timesteps. By analyzing layer-wise representation errors from a small calibration set, SmoothCache adaptively caches and reuses key features during inference. Our experiments demonstrate that SmoothCache achieves 8% to 71% speed up while maintaining or even improving generation quality across diverse modalities. We showcase its effectiveness on DiT-XL for image generation, Open-Sora for text-to-video, and Stable Audio Open for text-to-audio, highlighting its potential to enable real-time applications and broaden the accessibility of powerful DiT models.

Nov 20, 202427 min

Ep 92LLäMmlein: Compact and Competitive German-Only Language Models from Scratch

🤗 Paper Upvotes: 7 | cs.CL, cs.AI, cs.LG Authors: Jan Pfister, Julia Wunderle, Andreas Hotho Title: LLäMmlein: Compact and Competitive German-Only Language Models from Scratch Arxiv: http://arxiv.org/abs/2411.11171v1 Abstract: We create two German-only decoder models, LL\"aMmlein 120M and 1B, transparently from scratch and publish them, along with the training data, for the German NLP research community to use. The model training involved several key steps, including extensive data preprocessing, the creation of a custom German tokenizer, the training itself, as well as the evaluation of the final models on various benchmarks. Throughout the training process, multiple checkpoints were saved and analyzed using the SuperGLEBer benchmark to monitor the models' learning dynamics. Compared to state-of-the-art models on the SuperGLEBer benchmark, both LL\"aMmlein models performed competitively, consistently matching or surpassing models with similar parameter sizes. The results show that the models' quality scales with size as expected, but performance improvements on some tasks plateaued early, offering valuable insights into resource allocation for future model development.

Nov 20, 202422 min

Ep 91LLaVA-o1: Let Vision Language Models Reason Step-by-Step

🤗 Paper Upvotes: 64 | cs.CV Authors: Guowei Xu, Peng Jin, Li Hao, Yibing Song, Lichao Sun, Li Yuan Title: LLaVA-o1: Let Vision Language Models Reason Step-by-Step Arxiv: http://arxiv.org/abs/2411.10440v1 Abstract: Large language models have demonstrated substantial advancements in reasoning capabilities, particularly through inference-time scaling, as illustrated by models such as OpenAI's o1. However, current Vision-Language Models (VLMs) often struggle to perform systematic and structured reasoning, especially when handling complex visual question-answering tasks. In this work, we introduce LLaVA-o1, a novel VLM designed to conduct autonomous multistage reasoning. Unlike chain-of-thought prompting, LLaVA-o1 independently engages in sequential stages of summarization, visual interpretation, logical reasoning, and conclusion generation. This structured approach enables LLaVA-o1 to achieve marked improvements in precision on reasoning-intensive tasks. To accomplish this, we compile the LLaVA-o1-100k dataset, integrating samples from various visual question answering sources and providing structured reasoning annotations. Besides, we propose an inference-time stage-level beam search method, which enables effective inference-time scaling. Remarkably, with only 100k training samples and a simple yet effective inference time scaling method, LLaVA-o1 not only outperforms its base model by 8.9% on a wide range of multimodal reasoning benchmarks, but also surpasses the performance of larger and even closed-source models, such as Gemini-1.5-pro, GPT-4o-mini, and Llama-3.2-90B-Vision-Instruct.

Nov 19, 202425 min

Ep 90GaussianAnything: Interactive Point Cloud Latent Diffusion for 3D Generation

🤗 Paper Upvotes: 19 | cs.CV, cs.AI, cs.GR Authors: Yushi Lan, Shangchen Zhou, Zhaoyang Lyu, Fangzhou Hong, Shuai Yang, Bo Dai, Xingang Pan, Chen Change Loy Title: GaussianAnything: Interactive Point Cloud Latent Diffusion for 3D Generation Arxiv: http://arxiv.org/abs/2411.08033v1 Abstract: While 3D content generation has advanced significantly, existing methods still face challenges with input formats, latent space design, and output representations. This paper introduces a novel 3D generation framework that addresses these challenges, offering scalable, high-quality 3D generation with an interactive Point Cloud-structured Latent space. Our framework employs a Variational Autoencoder (VAE) with multi-view posed RGB-D(epth)-N(ormal) renderings as input, using a unique latent space design that preserves 3D shape information, and incorporates a cascaded latent diffusion model for improved shape-texture disentanglement. The proposed method, GaussianAnything, supports multi-modal conditional 3D generation, allowing for point cloud, caption, and single/multi-view image inputs. Notably, the newly proposed latent space naturally enables geometry-texture disentanglement, thus allowing 3D-aware editing. Experimental results demonstrate the effectiveness of our approach on multiple datasets, outperforming existing methods in both text- and image-conditioned 3D generation.

Nov 19, 202424 min

Ep 89Xmodel-1.5: An 1B-scale Multilingual LLM

🤗 Paper Upvotes: 7 | cs.CL Authors: Wang Qun, Liu Yang, Lin Qingquan, Jiang Ling Title: Xmodel-1.5: An 1B-scale Multilingual LLM Arxiv: http://arxiv.org/abs/2411.10083v1 Abstract: We introduce Xmodel-1.5, a novel 1-billion-parameter multilingual large model pretrained on approximately 2 trillion tokens. The model demonstrates strong performance across several languages, with particularly notable results in Thai, Arabic, and French, alongside its effectiveness in Chinese and English. In addition, we contribute to the research community by releasing a Thai evaluation dataset, which includes hundreds of questions annotated by students from Chulalongkorn University's School of Integrated Innovation. While the results are promising, we acknowledge that there is still room for improvement. We hope this work advances ongoing efforts in multilingual AI research and promotes better cross-linguistic understanding in various natural language processing tasks. Our models and code are publicly available on GitHub at https://github.com/XiaoduoAILab/XmodelLM.

Nov 19, 202420 min

Ep 88LLaMA-Mesh: Unifying 3D Mesh Generation with Language Models

🤗 Paper Upvotes: 32 | cs.LG, cs.AI, cs.CL, cs.CV, 68T05, I.3.5; I.2.10; I.2.6 Authors: Zhengyi Wang, Jonathan Lorraine, Yikai Wang, Hang Su, Jun Zhu, Sanja Fidler, Xiaohui Zeng Title: LLaMA-Mesh: Unifying 3D Mesh Generation with Language Models Arxiv: http://arxiv.org/abs/2411.09595v1 Abstract: This work explores expanding the capabilities of large language models (LLMs) pretrained on text to generate 3D meshes within a unified model. This offers key advantages of (1) leveraging spatial knowledge already embedded in LLMs, derived from textual sources like 3D tutorials, and (2) enabling conversational 3D generation and mesh understanding. A primary challenge is effectively tokenizing 3D mesh data into discrete tokens that LLMs can process seamlessly. To address this, we introduce LLaMA-Mesh, a novel approach that represents the vertex coordinates and face definitions of 3D meshes as plain text, allowing direct integration with LLMs without expanding the vocabulary. We construct a supervised fine-tuning (SFT) dataset enabling pretrained LLMs to (1) generate 3D meshes from text prompts, (2) produce interleaved text and 3D mesh outputs as required, and (3) understand and interpret 3D meshes. Our work is the first to demonstrate that LLMs can be fine-tuned to acquire complex spatial knowledge for 3D mesh generation in a text-based format, effectively unifying the 3D and text modalities. LLaMA-Mesh achieves mesh generation quality on par with models trained from scratch while maintaining strong text generation performance.

Nov 16, 202423 min

Ep 87MagicQuill: An Intelligent Interactive Image Editing System

🤗 Paper Upvotes: 31 | cs.CV Authors: Zichen Liu, Yue Yu, Hao Ouyang, Qiuyu Wang, Ka Leong Cheng, Wen Wang, Zhiheng Liu, Qifeng Chen, Yujun Shen Title: MagicQuill: An Intelligent Interactive Image Editing System Arxiv: http://arxiv.org/abs/2411.09703v1 Abstract: Image editing involves a variety of complex tasks and requires efficient and precise manipulation techniques. In this paper, we present MagicQuill, an integrated image editing system that enables swift actualization of creative ideas. Our system features a streamlined yet functionally robust interface, allowing for the articulation of editing operations (e.g., inserting elements, erasing objects, altering color) with minimal input. These interactions are monitored by a multimodal large language model (MLLM) to anticipate editing intentions in real time, bypassing the need for explicit prompt entry. Finally, we apply a powerful diffusion prior, enhanced by a carefully learned two-branch plug-in module, to process editing requests with precise control. Experimental results demonstrate the effectiveness of MagicQuill in achieving high-quality image edits. Please visit https://magic-quill.github.io to try out our system.

Nov 16, 202420 min

Ep 86Cut Your Losses in Large-Vocabulary Language Models

🤗 Paper Upvotes: 15 | cs.LG, cs.CL Authors: Erik Wijmans, Brody Huval, Alexander Hertzberg, Vladlen Koltun, Philipp Krähenbühl Title: Cut Your Losses in Large-Vocabulary Language Models Arxiv: http://arxiv.org/abs/2411.09009v1 Abstract: As language models grow ever larger, so do their vocabularies. This has shifted the memory footprint of LLMs during training disproportionately to one single layer: the cross-entropy in the loss computation. Cross-entropy builds up a logit matrix with entries for each pair of input tokens and vocabulary items and, for small models, consumes an order of magnitude more memory than the rest of the LLM combined. We propose Cut Cross-Entropy (CCE), a method that computes the cross-entropy loss without materializing the logits for all tokens into global memory. Rather, CCE only computes the logit for the correct token and evaluates the log-sum-exp over all logits on the fly. We implement a custom kernel that performs the matrix multiplications and the log-sum-exp reduction over the vocabulary in flash memory, making global memory consumption for the cross-entropy computation negligible. This has a dramatic effect. Taking the Gemma 2 (2B) model as an example, CCE reduces the memory footprint of the loss computation from 24 GB to 1 MB, and the total training-time memory consumption of the classifier head from 28 GB to 1 GB. To improve the throughput of CCE, we leverage the inherent sparsity of softmax and propose to skip elements of the gradient computation that have a negligible (i.e., below numerical precision) contribution to the gradient. Experiments demonstrate that the dramatic reduction in memory consumption is accomplished without sacrificing training speed or convergence.

Nov 16, 202420 min

Ep 85ClinicalBench: Can LLMs Beat Traditional ML Models in Clinical Prediction?

🤗 Paper Upvotes: 9 | cs.CL Authors: Canyu Chen, Jian Yu, Shan Chen, Che Liu, Zhongwei Wan, Danielle Bitterman, Fei Wang, Kai Shu Title: ClinicalBench: Can LLMs Beat Traditional ML Models in Clinical Prediction? Arxiv: http://arxiv.org/abs/2411.06469v1 Abstract: Large Language Models (LLMs) hold great promise to revolutionize current clinical systems for their superior capacities on medical text processing tasks and medical licensing exams. Meanwhile, traditional ML models such as SVM and XGBoost have still been mainly adopted in clinical prediction tasks. An emerging question is Can LLMs beat traditional ML models in clinical prediction? Thus, we build a new benchmark ClinicalBench to comprehensively study the clinical predictive modeling capacities of both general-purpose and medical LLMs, and compare them with traditional ML models. ClinicalBench embraces three common clinical prediction tasks, two databases, 14 general-purpose LLMs, 8 medical LLMs, and 11 traditional ML models. Through extensive empirical investigation, we discover that both general-purpose and medical LLMs, even with different model scales, diverse prompting or fine-tuning strategies, still cannot beat traditional ML models in clinical prediction yet, shedding light on their potential deficiency in clinical reasoning and decision-making. We call for caution when practitioners adopt LLMs in clinical applications. ClinicalBench can be utilized to bridge the gap between LLMs' development for healthcare and real-world clinical practice.

Nov 16, 202423 min

Ep 84Sharingan: Extract User Action Sequence from Desktop Recordings

🤗 Paper Upvotes: 3 | cs.CV, cs.AI Authors: Yanting Chen, Yi Ren, Xiaoting Qin, Jue Zhang, Kehong Yuan, Lu Han, Qingwei Lin, Dongmei Zhang, Saravan Rajmohan, Qi Zhang Title: Sharingan: Extract User Action Sequence from Desktop Recordings Arxiv: http://arxiv.org/abs/2411.08768v1 Abstract: Video recordings of user activities, particularly desktop recordings, offer a rich source of data for understanding user behaviors and automating processes. However, despite advancements in Vision-Language Models (VLMs) and their increasing use in video analysis, extracting user actions from desktop recordings remains an underexplored area. This paper addresses this gap by proposing two novel VLM-based methods for user action extraction: the Direct Frame-Based Approach (DF), which inputs sampled frames directly into VLMs, and the Differential Frame-Based Approach (DiffF), which incorporates explicit frame differences detected via computer vision techniques. We evaluate these methods using a basic self-curated dataset and an advanced benchmark adapted from prior work. Our results show that the DF approach achieves an accuracy of 70% to 80% in identifying user actions, with the extracted action sequences being re-playable though Robotic Process Automation. We find that while VLMs show potential, incorporating explicit UI changes can degrade performance, making the DF approach more reliable. This work represents the first application of VLMs for extracting user action sequences from desktop recordings, contributing new methods, benchmarks, and insights for future research.

Nov 16, 202422 min

Ep 83Hermes: A Large Language Model Framework on the Journey to Autonomous Networks

🤗 Paper Upvotes: 2 | cs.AI, cs.NI Authors: Fadhel Ayed, Ali Maatouk, Nicola Piovesan, Antonio De Domenico, Merouane Debbah, Zhi-Quan Luo Title: Hermes: A Large Language Model Framework on the Journey to Autonomous Networks Arxiv: http://arxiv.org/abs/2411.06490v1 Abstract: The drive toward automating cellular network operations has grown with the increasing complexity of these systems. Despite advancements, full autonomy currently remains out of reach due to reliance on human intervention for modeling network behaviors and defining policies to meet target requirements. Network Digital Twins (NDTs) have shown promise in enhancing network intelligence, but the successful implementation of this technology is constrained by use case-specific architectures, limiting its role in advancing network autonomy. A more capable network intelligence, or "telecommunications brain", is needed to enable seamless, autonomous management of cellular network. Large Language Models (LLMs) have emerged as potential enablers for this vision but face challenges in network modeling, especially in reasoning and handling diverse data types. To address these gaps, we introduce Hermes, a chain of LLM agents that uses "blueprints" for constructing NDT instances through structured and explainable logical steps. Hermes allows automatic, reliable, and accurate network modeling of diverse use cases and configurations, thus marking progress toward fully autonomous network operations.

Nov 16, 202422 min

Ep 82Inconsistencies In Consistency Models: Better ODE Solving Does Not Imply Better Samples

🤗 Paper Upvotes: 2 | cs.LG, cs.AI Authors: Noël Vouitsis, Rasa Hosseinzadeh, Brendan Leigh Ross, Valentin Villecroze, Satya Krishna Gorti, Jesse C. Cresswell, Gabriel Loaiza-Ganem Title: Inconsistencies In Consistency Models: Better ODE Solving Does Not Imply Better Samples Arxiv: http://arxiv.org/abs/2411.08954v1 Abstract: Although diffusion models can generate remarkably high-quality samples, they are intrinsically bottlenecked by their expensive iterative sampling procedure. Consistency models (CMs) have recently emerged as a promising diffusion model distillation method, reducing the cost of sampling by generating high-fidelity samples in just a few iterations. Consistency model distillation aims to solve the probability flow ordinary differential equation (ODE) defined by an existing diffusion model. CMs are not directly trained to minimize error against an ODE solver, rather they use a more computationally tractable objective. As a way to study how effectively CMs solve the probability flow ODE, and the effect that any induced error has on the quality of generated samples, we introduce Direct CMs, which \textit{directly} minimize this error. Intriguingly, we find that Direct CMs reduce the ODE solving error compared to CMs but also result in significantly worse sample quality, calling into question why exactly CMs work well in the first place. Full code is available at: https://github.com/layer6ai-labs/direct-cms.

Nov 16, 202422 min

Ep 81Direct Preference Optimization Using Sparse Feature-Level Constraints

🤗 Paper Upvotes: 10 | cs.AI, cs.CL Authors: Qingyu Yin, Chak Tou Leong, Hongbo Zhang, Minjun Zhu, Hanqi Yan, Qiang Zhang, Yulan He, Wenjie Li, Jun Wang, Yue Zhang, Linyi Yang Title: Direct Preference Optimization Using Sparse Feature-Level Constraints Arxiv: http://arxiv.org/abs/2411.07618v1 Abstract: The alignment of large language models (LLMs) with human preferences remains a key challenge. While post-training techniques like Reinforcement Learning from Human Feedback (RLHF) and Direct Preference Optimization (DPO) have achieved notable success, they often introduce computational inefficiencies and training instability. In this paper, we propose Feature-level constrained Preference Optimization (FPO), a novel method designed to simplify the alignment process while ensuring stability. FPO leverages pre-trained Sparse Autoencoders (SAEs) and introduces feature-level constraints, allowing for efficient, sparsity-enforced alignment. Our approach enjoys efficiency by using sparse features activated in a well-trained sparse autoencoder and the quality of sequential KL divergence by using the feature-level offline reference. Experimental results on benchmark datasets demonstrate that FPO achieves a 5.08% absolute improvement in win rate with much lower computational cost compared to state-of-the-art baselines, making it a promising solution for efficient and controllable LLM alignments.

Nov 15, 202421 min

Ep 80CamemBERT 2.0: A Smarter French Language Model Aged to Perfection

🤗 Paper Upvotes: 8 | cs.CL Authors: Wissam Antoun, Francis Kulumba, Rian Touchent, Éric de la Clergerie, Benoît Sagot, Djamé Seddah Title: CamemBERT 2.0: A Smarter French Language Model Aged to Perfection Arxiv: http://arxiv.org/abs/2411.08868v1 Abstract: French language models, such as CamemBERT, have been widely adopted across industries for natural language processing (NLP) tasks, with models like CamemBERT seeing over 4 million downloads per month. However, these models face challenges due to temporal concept drift, where outdated training data leads to a decline in performance, especially when encountering new topics and terminology. This issue emphasizes the need for updated models that reflect current linguistic trends. In this paper, we introduce two new versions of the CamemBERT base model-CamemBERTav2 and CamemBERTv2-designed to address these challenges. CamemBERTav2 is based on the DeBERTaV3 architecture and makes use of the Replaced Token Detection (RTD) objective for better contextual understanding, while CamemBERTv2 is built on RoBERTa, which uses the Masked Language Modeling (MLM) objective. Both models are trained on a significantly larger and more recent dataset with longer context length and an updated tokenizer that enhances tokenization performance for French. We evaluate the performance of these models on both general-domain NLP tasks and domain-specific applications, such as medical field tasks, demonstrating their versatility and effectiveness across a range of use cases. Our results show that these updated models vastly outperform their predecessors, making them valuable tools for modern NLP systems. All our new models, as well as intermediate checkpoints, are made openly available on Huggingface.

Nov 15, 202424 min

Ep 79Can sparse autoencoders be used to decompose and interpret steering vectors?

🤗 Paper Upvotes: 6 | cs.LG, cs.AI, cs.CL Authors: Harry Mayne, Yushi Yang, Adam Mahdi Title: Can sparse autoencoders be used to decompose and interpret steering vectors? Arxiv: http://arxiv.org/abs/2411.08790v1 Abstract: Steering vectors are a promising approach to control the behaviour of large language models. However, their underlying mechanisms remain poorly understood. While sparse autoencoders (SAEs) may offer a potential method to interpret steering vectors, recent findings show that SAE-reconstructed vectors often lack the steering properties of the original vectors. This paper investigates why directly applying SAEs to steering vectors yields misleading decompositions, identifying two reasons: (1) steering vectors fall outside the input distribution for which SAEs are designed, and (2) steering vectors can have meaningful negative projections in feature directions, which SAEs are not designed to accommodate. These limitations hinder the direct use of SAEs for interpreting steering vectors.

Nov 15, 202421 min

Ep 78PerceiverS: A Multi-Scale Perceiver with Effective Segmentation for Long-Term Expressive Symbolic Music Generation

🤗 Paper Upvotes: 5 | cs.AI, cs.MM, cs.SD, eess.AS Authors: Yungang Yi, Weihua Li, Matthew Kuo, Quan Bai Title: PerceiverS: A Multi-Scale Perceiver with Effective Segmentation for Long-Term Expressive Symbolic Music Generation Arxiv: http://arxiv.org/abs/2411.08307v1 Abstract: Music generation has progressed significantly, especially in the domain of audio generation. However, generating symbolic music that is both long-structured and expressive remains a significant challenge. In this paper, we propose PerceiverS (Segmentation and Scale), a novel architecture designed to address this issue by leveraging both Effective Segmentation and Multi-Scale attention mechanisms. Our approach enhances symbolic music generation by simultaneously learning long-term structural dependencies and short-term expressive details. By combining cross-attention and self-attention in a Multi-Scale setting, PerceiverS captures long-range musical structure while preserving performance nuances. The proposed model, evaluated on datasets like Maestro, demonstrates improvements in generating coherent and diverse music with both structural consistency and expressive variation. The project demos and the generated music samples can be accessed through the link: https://perceivers.github.io.

Nov 15, 202418 min

Ep 77SAMPart3D: Segment Any Part in 3D Objects

🤗 Paper Upvotes: 18 | cs.CV Authors: Yunhan Yang, Yukun Huang, Yuan-Chen Guo, Liangjun Lu, Xiaoyang Wu, Edmund Y. Lam, Yan-Pei Cao, Xihui Liu Title: SAMPart3D: Segment Any Part in 3D Objects Arxiv: http://arxiv.org/abs/2411.07184v1 Abstract: 3D part segmentation is a crucial and challenging task in 3D perception, playing a vital role in applications such as robotics, 3D generation, and 3D editing. Recent methods harness the powerful Vision Language Models (VLMs) for 2D-to-3D knowledge distillation, achieving zero-shot 3D part segmentation. However, these methods are limited by their reliance on text prompts, which restricts the scalability to large-scale unlabeled datasets and the flexibility in handling part ambiguities. In this work, we introduce SAMPart3D, a scalable zero-shot 3D part segmentation framework that segments any 3D object into semantic parts at multiple granularities, without requiring predefined part label sets as text prompts. For scalability, we use text-agnostic vision foundation models to distill a 3D feature extraction backbone, allowing scaling to large unlabeled 3D datasets to learn rich 3D priors. For flexibility, we distill scale-conditioned part-aware 3D features for 3D part segmentation at multiple granularities. Once the segmented parts are obtained from the scale-conditioned part-aware 3D features, we use VLMs to assign semantic labels to each part based on the multi-view renderings. Compared to previous methods, our SAMPart3D can scale to the recent large-scale 3D object dataset Objaverse and handle complex, non-ordinary objects. Additionally, we contribute a new 3D part segmentation benchmark to address the lack of diversity and complexity of objects and parts in existing benchmarks. Experiments show that our SAMPart3D significantly outperforms existing zero-shot 3D part segmentation methods, and can facilitate various applications such as part-level editing and interactive segmentation.

Nov 14, 202420 min

Ep 76JanusFlow: Harmonizing Autoregression and Rectified Flow for Unified Multimodal Understanding and Generation

🤗 Paper Upvotes: 14 | cs.CV, cs.AI, cs.CL Authors: Yiyang Ma, Xingchao Liu, Xiaokang Chen, Wen Liu, Chengyue Wu, Zhiyu Wu, Zizheng Pan, Zhenda Xie, Haowei Zhang, Xingkai yu, Liang Zhao, Yisong Wang, Jiaying Liu, Chong Ruan Title: JanusFlow: Harmonizing Autoregression and Rectified Flow for Unified Multimodal Understanding and Generation Arxiv: http://arxiv.org/abs/2411.07975v1 Abstract: We present JanusFlow, a powerful framework that unifies image understanding and generation in a single model. JanusFlow introduces a minimalist architecture that integrates autoregressive language models with rectified flow, a state-of-the-art method in generative modeling. Our key finding demonstrates that rectified flow can be straightforwardly trained within the large language model framework, eliminating the need for complex architectural modifications. To further improve the performance of our unified model, we adopt two key strategies: (i) decoupling the understanding and generation encoders, and (ii) aligning their representations during unified training. Extensive experiments show that JanusFlow achieves comparable or superior performance to specialized models in their respective domains, while significantly outperforming existing unified approaches across standard benchmarks. This work represents a step toward more efficient and versatile vision-language models.

Nov 14, 202422 min

Ep 75Stronger Models are NOT Stronger Teachers for Instruction Tuning

🤗 Paper Upvotes: 13 | cs.AI, cs.CL Authors: Zhangchen Xu, Fengqing Jiang, Luyao Niu, Bill Yuchen Lin, Radha Poovendran Title: Stronger Models are NOT Stronger Teachers for Instruction Tuning Arxiv: http://arxiv.org/abs/2411.07133v2 Abstract: Instruction tuning has been widely adopted to ensure large language models (LLMs) follow user instructions effectively. The resulting instruction-following capabilities of LLMs heavily rely on the instruction datasets used for tuning. Recently, synthetic instruction datasets have emerged as an economically viable solution to provide LLMs diverse and high-quality instructions. However, existing approaches typically assume that larger or stronger models are stronger teachers for instruction tuning, and hence simply adopt these models as response generators to the synthetic instructions. In this paper, we challenge this commonly-adopted assumption. Our extensive experiments across five base models and twenty response generators reveal that larger and stronger models are not necessarily stronger teachers of smaller models. We refer to this phenomenon as the Larger Models' Paradox. We observe that existing metrics cannot precisely predict the effectiveness of response generators since they ignore the compatibility between teachers and base models being fine-tuned. We thus develop a novel metric, named as Compatibility-Adjusted Reward (CAR) to measure the effectiveness of response generators. Our experiments across five base models demonstrate that CAR outperforms almost all baselines.

Nov 14, 202427 min

Ep 74BLIP3-KALE: Knowledge Augmented Large-Scale Dense Captions

🤗 Paper Upvotes: 11 | cs.CV, cs.AI Authors: Anas Awadalla, Le Xue, Manli Shu, An Yan, Jun Wang, Senthil Purushwalkam, Sheng Shen, Hannah Lee, Oscar Lo, Jae Sung Park, Etash Guha, Silvio Savarese, Ludwig Schmidt, Yejin Choi, Caiming Xiong, Ran Xu Title: BLIP3-KALE: Knowledge Augmented Large-Scale Dense Captions Arxiv: http://arxiv.org/abs/2411.07461v1 Abstract: We introduce BLIP3-KALE, a dataset of 218 million image-text pairs that bridges the gap between descriptive synthetic captions and factual web-scale alt-text. KALE augments synthetic dense image captions with web-scale alt-text to generate factually grounded image captions. Our two-stage approach leverages large vision-language models and language models to create knowledge-augmented captions, which are then used to train a specialized VLM for scaling up the dataset. We train vision-language models on KALE and demonstrate improvements on vision-language tasks. Our experiments show the utility of KALE for training more capable and knowledgeable multimodal models. We release the KALE dataset at https://huggingface.co/datasets/Salesforce/blip3-kale

Nov 14, 202420 min

Ep 73Scaling Properties of Diffusion Models for Perceptual Tasks

🤗 Paper Upvotes: 7 | cs.CV, cs.AI Authors: Rahul Ravishankar, Zeeshan Patel, Jathushan Rajasegaran, Jitendra Malik Title: Scaling Properties of Diffusion Models for Perceptual Tasks Arxiv: http://arxiv.org/abs/2411.08034v2 Abstract: In this paper, we argue that iterative computation with diffusion models offers a powerful paradigm for not only generation but also visual perception tasks. We unify tasks such as depth estimation, optical flow, and amodal segmentation under the framework of image-to-image translation, and show how diffusion models benefit from scaling training and test-time compute for these perceptual tasks. Through a careful analysis of these scaling properties, we formulate compute-optimal training and inference recipes to scale diffusion models for visual perception tasks. Our models achieve competitive performance to state-of-the-art methods using significantly less data and compute. To access our code and models, see https://scaling-diffusion-perception.github.io .

Nov 14, 202425 min

Ep 72Wavelet Latent Diffusion (Wala): Billion-Parameter 3D Generative Model with Compact Wavelet Encodings

🤗 Paper Upvotes: 5 | cs.CV, cs.AI, cs.LG Authors: Aditya Sanghi, Aliasghar Khani, Pradyumna Reddy, Arianna Rampini, Derek Cheung, Kamal Rahimi Malekshan, Kanika Madan, Hooman Shayani Title: Wavelet Latent Diffusion (Wala): Billion-Parameter 3D Generative Model with Compact Wavelet Encodings Arxiv: http://arxiv.org/abs/2411.08017v1 Abstract: Large-scale 3D generative models require substantial computational resources yet often fall short in capturing fine details and complex geometries at high resolutions. We attribute this limitation to the inefficiency of current representations, which lack the compactness required to model the generative models effectively. To address this, we introduce a novel approach called Wavelet Latent Diffusion, or WaLa, that encodes 3D shapes into wavelet-based, compact latent encodings. Specifically, we compress a $256^3$ signed distance field into a $12^3 \times 4$ latent grid, achieving an impressive 2427x compression ratio with minimal loss of detail. This high level of compression allows our method to efficiently train large-scale generative networks without increasing the inference time. Our models, both conditional and unconditional, contain approximately one billion parameters and successfully generate high-quality 3D shapes at $256^3$ resolution. Moreover, WaLa offers rapid inference, producing shapes within two to four seconds depending on the condition, despite the model's scale. We demonstrate state-of-the-art performance across multiple datasets, with significant improvements in generation quality, diversity, and computational efficiency. We open-source our code and, to the best of our knowledge, release the largest pretrained 3D generative models across different modalities.

Nov 14, 202422 min

Ep 71Add-it: Training-Free Object Insertion in Images With Pretrained Diffusion Models

🤗 Paper Upvotes: 44 | cs.CV, cs.AI, cs.GR, cs.LG Authors: Yoad Tewel, Rinon Gal, Dvir Samuel, Yuval Atzmon, Lior Wolf, Gal Chechik Title: Add-it: Training-Free Object Insertion in Images With Pretrained Diffusion Models Arxiv: http://arxiv.org/abs/2411.07232v2 Abstract: Adding Object into images based on text instructions is a challenging task in semantic image editing, requiring a balance between preserving the original scene and seamlessly integrating the new object in a fitting location. Despite extensive efforts, existing models often struggle with this balance, particularly with finding a natural location for adding an object in complex scenes. We introduce Add-it, a training-free approach that extends diffusion models' attention mechanisms to incorporate information from three key sources: the scene image, the text prompt, and the generated image itself. Our weighted extended-attention mechanism maintains structural consistency and fine details while ensuring natural object placement. Without task-specific fine-tuning, Add-it achieves state-of-the-art results on both real and generated image insertion benchmarks, including our newly constructed "Additing Affordance Benchmark" for evaluating object placement plausibility, outperforming supervised methods. Human evaluations show that Add-it is preferred in over 80% of cases, and it also demonstrates improvements in various automated metrics.

Nov 13, 202423 min

Ep 70OmniEdit: Building Image Editing Generalist Models Through Specialist Supervision

🤗 Paper Upvotes: 39 | cs.CV, cs.AI Authors: Cong Wei, Zheyang Xiong, Weiming Ren, Xinrun Du, Ge Zhang, Wenhu Chen Title: OmniEdit: Building Image Editing Generalist Models Through Specialist Supervision Arxiv: http://arxiv.org/abs/2411.07199v1 Abstract: Instruction-guided image editing methods have demonstrated significant potential by training diffusion models on automatically synthesized or manually annotated image editing pairs. However, these methods remain far from practical, real-life applications. We identify three primary challenges contributing to this gap. Firstly, existing models have limited editing skills due to the biased synthesis process. Secondly, these methods are trained with datasets with a high volume of noise and artifacts. This is due to the application of simple filtering methods like CLIP-score. Thirdly, all these datasets are restricted to a single low resolution and fixed aspect ratio, limiting the versatility to handle real-world use cases. In this paper, we present \omniedit, which is an omnipotent editor to handle seven different image editing tasks with any aspect ratio seamlessly. Our contribution is in four folds: (1) \omniedit is trained by utilizing the supervision from seven different specialist models to ensure task coverage. (2) we utilize importance sampling based on the scores provided by large multimodal models (like GPT-4o) instead of CLIP-score to improve the data quality. (3) we propose a new editing architecture called EditNet to greatly boost the editing success rate, (4) we provide images with different aspect ratios to ensure that our model can handle any image in the wild. We have curated a test set containing images of different aspect ratios, accompanied by diverse instructions to cover different tasks. Both automatic evaluation and human evaluations demonstrate that \omniedit can significantly outperform all the existing models. Our code, dataset and model will be available at \url{https://tiger-ai-lab.github.io/OmniEdit/}

Nov 13, 202419 min

Ep 69Chinese SimpleQA: A Chinese Factuality Evaluation for Large Language Models

🤗 Paper Upvotes: 30 | cs.CL Authors: Yancheng He, Shilong Li, Jiaheng Liu, Yingshui Tan, Hui Huang, Weixun Wang, Xingyuan Bu, Hangyu Guo, Chengwei Hu, Boren Zheng, Xuepeng Liu, Dekai Sun, Wenbo Su, Bo Zheng Title: Chinese SimpleQA: A Chinese Factuality Evaluation for Large Language Models Arxiv: http://arxiv.org/abs/2411.07140v1 Abstract: New LLM evaluation benchmarks are important to align with the rapid development of Large Language Models (LLMs). In this work, we present Chinese SimpleQA, the first comprehensive Chinese benchmark to evaluate the factuality ability of language models to answer short questions, and Chinese SimpleQA mainly has five properties (i.e., Chinese, Diverse, High-quality, Static, Easy-to-evaluate). Specifically, first, we focus on the Chinese language over 6 major topics with 99 diverse subtopics. Second, we conduct a comprehensive quality control process to achieve high-quality questions and answers, where the reference answers are static and cannot be changed over time. Third, following SimpleQA, the questions and answers are very short, and the grading process is easy-to-evaluate based on OpenAI API. Based on Chinese SimpleQA, we perform a comprehensive evaluation on the factuality abilities of existing LLMs. Finally, we hope that Chinese SimpleQA could guide the developers to better understand the Chinese factuality abilities of their models and facilitate the growth of foundation models.

Nov 13, 202421 min

Ep 68M-Longdoc: A Benchmark For Multimodal Super-Long Document Understanding And A Retrieval-Aware Tuning Framework

🤗 Paper Upvotes: 28 | cs.CL Authors: Yew Ken Chia, Liying Cheng, Hou Pong Chan, Chaoqun Liu, Maojia Song, Sharifah Mahani Aljunied, Soujanya Poria, Lidong Bing Title: M-Longdoc: A Benchmark For Multimodal Super-Long Document Understanding And A Retrieval-Aware Tuning Framework Arxiv: http://arxiv.org/abs/2411.06176v1 Abstract: The ability to understand and answer questions over documents can be useful in many business and practical applications. However, documents often contain lengthy and diverse multimodal contents such as texts, figures, and tables, which are very time-consuming for humans to read thoroughly. Hence, there is an urgent need to develop effective and automated methods to aid humans in this task. In this work, we introduce M-LongDoc, a benchmark of 851 samples, and an automated framework to evaluate the performance of large multimodal models. We further propose a retrieval-aware tuning approach for efficient and effective multimodal document reading. Compared to existing works, our benchmark consists of more recent and lengthy documents with hundreds of pages, while also requiring open-ended solutions and not just extractive answers. To our knowledge, our training framework is the first to directly address the retrieval setting for multimodal long documents. To enable tuning open-source models, we construct a training corpus in a fully automatic manner for the question-answering task over such documents. Experiments show that our tuning approach achieves a relative improvement of 4.6% for the correctness of model responses, compared to the baseline open-source models. Our data, code, and models are available at https://multimodal-documents.github.io.

Nov 13, 202420 min

Ep 67Edify Image: High-Quality Image Generation with Pixel Space Laplacian Diffusion Models

🤗 Paper Upvotes: 21 | cs.CV, cs.LG Authors: NVIDIA, :, Yuval Atzmon, Maciej Bala, Yogesh Balaji, Tiffany Cai, Yin Cui, Jiaojiao Fan, Yunhao Ge, Siddharth Gururani, Jacob Huffman, Ronald Isaac, Pooya Jannaty, Tero Karras, Grace Lam, J. P. Lewis, Aaron Licata, Yen-Chen Lin, Ming-Yu Liu, Qianli Ma, Arun Mallya, Ashlee Martino-Tarr, Doug Mendez, Seungjun Nah, Chris Pruett, Fitsum Reda, Jiaming Song, Ting-Chun Wang, Fangyin Wei, Xiaohui Zeng, Yu Zeng, Qinsheng Zhang Title: Edify Image: High-Quality Image Generation with Pixel Space Laplacian Diffusion Models Arxiv: http://arxiv.org/abs/2411.07126v1 Abstract: We introduce Edify Image, a family of diffusion models capable of generating photorealistic image content with pixel-perfect accuracy. Edify Image utilizes cascaded pixel-space diffusion models trained using a novel Laplacian diffusion process, in which image signals at different frequency bands are attenuated at varying rates. Edify Image supports a wide range of applications, including text-to-image synthesis, 4K upsampling, ControlNets, 360 HDR panorama generation, and finetuning for image customization.

Nov 13, 202424 min