Total relevant papers: 40
Paper selection prompt and criteria at the bottom
Table of contents with paper titles:
AMiD: Knowledge Distillation for LLMs with $\alpha$-mixture Assistant Distribution Authors: Donghyeok Shin, Yeongmin Kim, Suhyeon Jo, Byeonghu Na, Il-Chul Moon
Localist LLMs with Recruitment Learning Authors: Joachim Diederich
MGTS-Net: Exploring Graph-Enhanced Multimodal Fusion for Augmented Time Series Forecasting Authors: Shule Hao, Junpeng Bao, Wenli Li
QSVD: Efficient Low-rank Approximation for Unified Query-Key-Value Weight Compression in Low-Precision Vision-Language Models Authors: Yutong Wang, Haiyu Wang, Sai Qian Zhang
Learning from Mistakes: Enhancing Harmful Meme Detection via Misjudgment Risk Patterns Authors: Wenshuo Wang, Ziyou Jiang, Junjie Wang, Mingyang Li, Jie Huang, Yuekai Huang, Zhiyuan Chang, Feiyan Duan, Qing Wang
End-to-end Listen, Look, Speak and Act Authors: Siyin Wang, Wenyi Yu, Xianzhao Chen, Xiaohai Tian, Jun Zhang, Lu Lu, Chao Zhang
Stratos: An End-to-End Distillation Pipeline for Customized LLMs under Distributed Cloud Environments Authors: Ziming Dai, Tuo Zhang, Fei Gao, Xingyi Cai, Xiaofei Wang, Cheng Zhang, Wenyu Wang, Chengjie Zang
ProtoMol: Enhancing Molecular Property Prediction via Prototype-Guided Multimodal Learning Authors: Yingxu Wang, Kunyu Zhang, Jiaxin Huang, Nan Yin, Siwei Liu, Eran Segal
Learning to play: A Multimodal Agent for 3D Game-Play Authors: Yuguang Yue, Irakli Salia, Samuel Hunt, Christopher Green, Wenzhe Shi, Jonathan J Hunt
VisuoAlign: Safety Alignment of LVLMs with Multimodal Tree Search Authors: MingSheng Li, Guangze Zhao, Sichen Liu
SimpleVSF: VLM-Scoring Fusion for Trajectory Prediction of End-to-End Autonomous Driving Authors: Peiru Zheng, Yun Zhao, Zhan Gong, Hong Zhu, Shaohua Wu
Vector Quantization in the Brain: Grid-like Codes in World Models Authors: Xiangyuan Peng, Xingsi Dong, Si Wu
MoS-VLA: A Vision-Language-Action Model with One-Shot Skill Adaptation Authors: Ruihan Zhao, Tyler Ingebrand, Sandeep Chinchali, Ufuk Topcu
Continual Knowledge Consolidation LORA for Domain Incremental Learning Authors: Naeem Paeedeh, Mahardhika Pratama, Weiping Ding, Jimmy Cao, Wolfgang Mayer, Ryszard Kowalczyk
EEschematic: Multimodal-LLM Based AI Agent for Schematic Generation of Analog Circuit Authors: Chang Liu, Danial Chitnis
Graph4MM: Weaving Multimodal Learning with Structural Information Authors: Xuying Ning, Dongqi Fu, Tianxin Wei, Wujiang Xu, Jingrui He
Long Exposure: Accelerating Parameter-Efficient Fine-Tuning for LLMs under Shadowy Sparsity Authors: Tuowei Wang, Kun Li, Zixu Hao, Donglin Bai, Ju Ren, Yaoxue Zhang, Ting Cao, Mao Yang
SOLE: Hardware-Software Co-design of Softmax and LayerNorm for Efficient Transformer Inference Authors: Wenxun Wang, Shuchang Zhou, Wenyu Sun, Peiqin Sun, Yongpan Liu
AMS-QUANT: Adaptive Mantissa Sharing for Floating-point Quantization Authors: Mengtao Lv, Ruiqi Zhu, Xinyu Wang, Yun Li
OmniVIC: A Self-Improving Variable Impedance Controller with Vision-Language In-Context Learning for Safe Robotic Manipulation Authors: Heng Zhang, Wei-Hsing Huang, Gokhan Solak, Arash Ajoudani
Compressing Many-Shots in In-Context Learning Authors: Devvrit Khatri, Pranamya Kulkarni, Nilesh Gupta, Yerram Varun, Liqian Peng, Jay Yagnik, Praneeth Netrapalli, Cho-Jui Hsieh, Alec Go, Inderjit S Dhillon, Aditya Kusupati, Prateek Jain
MIRAGE: Agentic Framework for Multimodal Misinformation Detection with Web-Grounded Reasoning Authors: Mir Nafis Sharear Shopnil, Sharad Duwal, Abhishek Tyagi, Adiba Mahbub Proma
See or Say Graphs: Agent-Driven Scalable Graph Understanding with Vision-Language Models Authors: Shuo Han, Yukun Cao, Zezhong Ding, Zengyi Gao, S Kevin Zhou, Xike Xie
BEACON: Bayesian Optimal Stopping for Efficient LLM Sampling Authors: Guangya Wan, Zixin Stephen Xu, Sasa Zorc, Manel Baucells, Mengxuan Hu, Hao Wang, Sheng Li
STABLE: Gated Continual Learning for Large Language Models Authors: William Hoy, Nurcin Celik
ELMM: Efficient Lightweight Multimodal Large Language Models for Multimodal Knowledge Graph Completion Authors: Wei Huang, Peining Li, Meiyu Liang, Xu Hou, Junping Du, Yingxia Shao, Guanhua Ye, Wu Liu, Kangkang Lu, Yang Yu
VAGEN: Reinforcing World Model Reasoning for Multi-Turn VLM Agents Authors: Kangrui Wang, Pingyue Zhang, Zihan Wang, Yaning Gao, Linjie Li, Qineng Wang, Hanyang Chen, Chi Wan, Yiping Lu, Zhengyuan Yang, Lijuan Wang, Ranjay Krishna, Jiajun Wu, Li Fei-Fei, Yejin Choi, Manling Li
CTR-LoRA: Curvature-Aware and Trust-Region Guided Low-Rank Adaptation for Large Language Models Authors: Zhuxuanzi Wang, Mingqiao Mo, Xi Xiao, Chen Liu, Chenrui Ma, Yunbei Zhang, Xiao Wang, Smita Krishnaswamy, Tianyang Wang
Input Domain Aware MoE: Decoupling Routing Decisions from Task Optimization in Mixture of Experts Authors: Yongxiang Hua, Haoyu Cao, Zhou Tao, Bocheng Li, Zihao Wu, Chaohu Liu, Linli Xu
Mixed-Precision Quantization for Language Models: Techniques and Prospects Authors: Mariam Rakka, Marios Fournarakis, Olga Krestinskaya, Jinane Bazzi, Khaled N. Salama, Fadi Kurdahi, Ahmed M. Eltawil, Mohammed E. Fouda
Modeling Expert Interactions in Sparse Mixture of Experts via Graph Structures Authors: Minh-Khoi Nguyen-Nhat, Rachel S. Y. Teo, Laziz Abdullaev, Maurice Mok, Viet-Hoang Tran, Tan Minh Nguyen
Do What You Say: Steering Vision-Language-Action Models via Runtime Reasoning-Action Alignment Verification Authors: Yilin Wu, Anqi Li, Tucker Hermans, Fabio Ramos, Andrea Bajcsy, Claudia P'erez-D'Arpino
Seeing but Not Believing: Probing the Disconnect Between Visual Attention and Answer Correctness in VLMs Authors: Zhining Liu, Ziyi Chen, Hui Liu, Chen Luo, Xianfeng Tang, Suhang Wang, Joy Zeng, Zhenwei Dai, Zhan Shi, Tianxin Wei, Benoit Dumoulin, Hanghang Tong
Expert Merging in Sparse Mixture of Experts with Nash Bargaining Authors: Dung V. Nguyen, Anh T. Nguyen, Minh H. Nguyen, Luc Q. Nguyen, Shiqi Jiang, Ethan Fetaya, Linh Duy Tran, Gal Chechik, Tan M. Nguyen
From Spatial to Actions: Grounding Vision-Language-Action Model in Spatial Foundation Priors Authors: Zhengshen Zhang, Hao Li, Yalun Dai, Zhengbang Zhu, Lei Zhou, Chenchen Liu, Dong Wang, Francis E. H. Tay, Sijin Chen, Ziwei Liu, Yuxiao Liu, Xinghang Li, Pan Zhou
Bitwidth-Specific Logarithmic Arithmetic for Future Hardware-Accelerated Training Authors: Hassan Hamad, Yuou Qiu, Peter A. Beerel, Keith M. Chugg
Bridging Embodiment Gaps: Deploying Vision-Language-Action Models on Soft Robots Authors: Haochen Su, Cristian Meo, Francesco Stella, Andrea Peirone, Kai Junge, Josie Hughes
Robobench: A Comprehensive Evaluation Benchmark for Multimodal Large Language Models as Embodied Brain Authors: Yulin Luo, Chun-Kai Fan, Menghang Dong, Jiayu Shi, Mengdi Zhao, Bo-Wen Zhang, Cheng Chi, Jiaming Liu, Gaole Dai, Rongyu Zhang, Ruichuan An, Kun Wu, Zhengping Che, Shaoxuan Xie, Guocai Yao, Zhongxia Zhao, Pengwei Wang, Guang Liu, Zhongyuan Wang, Tiejun Huang, Shanghang Zhang
Unbiased Gradient Low-Rank Projection Authors: Rui Pan, Yang Luo, Yuxing Liu, Yang You, Tong Zhang
MILES: Modality-Informed Learning Rate Scheduler for Balancing Multimodal Learning Authors: Alejandro Guerra-Manzanares, Farah E. Shamout
ArXiv ID: 2510.15982 Authors: Donghyeok Shin, Yeongmin Kim, Suhyeon Jo, Byeonghu Na, Il-Chul Moon
Abstract: arXiv:2510.15982v1 Announce Type: new Abstract: Autoregressive large language models (LLMs) have achieved remarkable improvement across many tasks but incur high computational and memory costs. Knowledge distillation (KD) mitigates this issue by transferring knowledge from a large teacher to a smaller student through distributional alignment. Previous studies have proposed various discrepancy metrics, but the capacity gap and training instability caused by near-zero probabilities, stemming from the high-dimensional output of LLMs, remain fundamental limitations. To overcome these challenges, several approaches implicitly or explicitly incorporating assistant distribution have recently been proposed. However, the past proposals of assistant distributions have been a fragmented approach without a systematic investigation of the interpolation path and the divergence. This paper proposes $\alpha$-mixture assistant distribution, a novel generalized family of assistant distributions, and $\alpha$-mixture distillation, coined AMiD, a unified framework for KD using the assistant distribution. The $\alpha$-mixture assistant distribution provides a continuous extension of the assistant distribution by introducing a new distribution design variable $\alpha$, which has been fixed in all previous approaches. Furthermore, AMiD generalizes the family of divergences used with the assistant distributions based on optimality, which has also been restricted in previous works. Through extensive experiments, we demonstrate that AMiD offers superior performance and training stability by leveraging a broader and theoretically grounded assistant distribution space.
ArXiv ID: 2510.17358 Authors: Joachim Diederich
Abstract: arXiv:2510.17358v1 Announce Type: new Abstract: We present a novel framework for training large language models with continuously adjustable internal representations that span the full spectrum from localist (interpretable, rule-based) to distributed (generalizable, efficient) encodings. The key innovations are (1) a locality dial, a tunable parameter that dynamically controls the degree of localization during both training and inference without requiring model retraining, (2) an information-theoretic recruitment mechanism that adaptively allocates semantic blocks as needed, eliminating the requirement for complete domain knowledge at initialization, and (3) a hierarchical recruitment framework that extends capacity allocation to entire specialized LLMs, enabling multi-granularity architectural adaptation. This is achieved through group sparsity penalties on attention mechanisms, information-theoretic anchor design, dynamic rule injection, and principled recruitment criteria based on penalized likelihood with explicit units. We provide rigorous mathematical results establishing explicit threshold conditions under which attention provably concentrates on semantically relevant blocks at stationary points, with exact bounds on attention entropy and pointer fidelity. The hierarchical recruitment mechanism provides convergence guarantees at both the block level (fine-grained, within-LLM) and the LLM level (coarse-grained, cross-domain), ensuring the system discovers semantic partitions that balance model complexity against data encoding efficiency. This framework enables practitioners to continuously interpolate between interpretable and high-performance modes while adapting architectural capacity at multiple granularities, supporting applications in regulated domains requiring both transparency and capability.
ArXiv ID: 2510.16350 Authors: Shule Hao, Junpeng Bao, Wenli Li
Abstract: arXiv:2510.16350v1 Announce Type: new Abstract: Recent research in time series forecasting has explored integrating multimodal features into models to improve accuracy. However, the accuracy of such methods is constrained by three key challenges: inadequate extraction of fine-grained temporal patterns, suboptimal integration of multimodal information, and limited adaptability to dynamic multi-scale features. To address these problems, we propose MGTS-Net, a Multimodal Graph-enhanced Network for Time Series forecasting. The model consists of three core components: (1) a Multimodal Feature Extraction layer (MFE), which optimizes feature encoders according to the characteristics of temporal, visual, and textual modalities to extract temporal features of fine-grained patterns; (2) a Multimodal Feature Fusion layer (MFF), which constructs a heterogeneous graph to model intra-modal temporal dependencies and cross-modal alignment relationships and dynamically aggregates multimodal knowledge; (3) a Multi-Scale Prediction layer (MSP), which adapts to multi-scale features by dynamically weighting and fusing the outputs of short-term, medium-term, and long-term predictors. Extensive experiments demonstrate that MGTS-Net exhibits excellent performance with light weight and high efficiency. Compared with other state-of-the-art baseline models, our method achieves superior performance, validating the superiority of the proposed methodology.
ArXiv ID: 2510.16292 Authors: Yutong Wang, Haiyu Wang, Sai Qian Zhang
Abstract: arXiv:2510.16292v1 Announce Type: new Abstract: Vision-Language Models (VLMs) are integral to tasks such as image captioning and visual question answering, but their high computational cost, driven by large memory footprints and processing time, limits their scalability and real-time applicability. In this work, we propose leveraging Singular-Value Decomposition (SVD) over the joint query (Q), key (K), and value (V) weight matrices to reduce KV cache size and computational overhead. We in addition introduce an efficient rank allocation strategy that dynamically adjusts the SVD rank based on its impact on VLM accuracy, achieving a significant reduction in both memory usage and computational cost. Finally, we extend this approach by applying quantization to both VLM weights and activations, resulting in a highly efficient VLM. Our method outperforms previous approaches that rely solely on quantization or SVD by achieving more than $10%$ accuracy improvement while consuming less hardware cost, making it better for real-time deployment on resource-constrained devices. We open source our code at \href{https://github.com/SAI-Lab-NYU/QSVD}{\texttt{https://github.com/SAI-Lab-NYU/QSVD}}.
ArXiv ID: 2510.15946 Authors: Wenshuo Wang, Ziyou Jiang, Junjie Wang, Mingyang Li, Jie Huang, Yuekai Huang, Zhiyuan Chang, Feiyan Duan, Qing Wang
Abstract: arXiv:2510.15946v1 Announce Type: new Abstract: Internet memes have emerged as a popular multimodal medium, yet they are increasingly weaponized to convey harmful opinions through subtle rhetorical devices like irony and metaphor. Existing detection approaches, including MLLM-based techniques, struggle with these implicit expressions, leading to frequent misjudgments. This paper introduces PatMD, a novel approach that improves harmful meme detection by learning from and proactively mitigating these potential misjudgment risks. Our core idea is to move beyond superficial content-level matching and instead identify the underlying misjudgment risk patterns, proactively guiding the MLLMs to avoid known misjudgment pitfalls. We first construct a knowledge base where each meme is deconstructed into a misjudgment risk pattern explaining why it might be misjudged, either overlooking harmful undertones (false negative) or overinterpreting benign content (false positive). For a given target meme, PatMD retrieves relevant patterns and utilizes them to dynamically guide the MLLM's reasoning. Experiments on a benchmark of 6,626 memes across 5 harmful detection tasks show that PatMD outperforms state-of-the-art baselines, achieving an average of 8.30% improvement in F1-score and 7.71% improvement in accuracy, demonstrating strong generalizability and improved detection capability of harmful memes.
ArXiv ID: 2510.16756 Authors: Siyin Wang, Wenyi Yu, Xianzhao Chen, Xiaohai Tian, Jun Zhang, Lu Lu, Chao Zhang
Abstract: arXiv:2510.16756v1 Announce Type: new Abstract: Human interaction is inherently multimodal and full-duplex: we listen while watching, speak while acting, and fluidly adapt to turn-taking and interruptions. Realizing these capabilities is essential for building models simulating humans. We present ELLSA (End-to-end Listen, Look, Speak and Act), which, to our knowledge, is the first full-duplex, end-to-end model that simultaneously perceives and generates across vision, text, speech, and action within a single architecture, enabling interaction patterns previously out of reach, yielding more natural, human-like behaviors. At its core is a novel SA-MoE architecture (Self-Attention Mixture-of-Experts) that routes each modality to specialized experts and fuses them through a unified attention backbone. This provides a generalizable solution for joint multimodal perception and concurrent generation, leveraging strong pre-trained components while enabling efficient modality integration and mitigating modality interference. On speech-interaction and robot-manipulation benchmarks, ELLSA matches modality-specific baselines, while uniquely supporting advanced multimodal and full-duplex behaviors such as dialogue and action turn-taking, defective instruction rejection, speaking-while-acting, context-grounded visual question answering, and action barge-ins. We contend that ELLSA represents a step toward more natural and general interactive intelligence, contributing to the broader pursuit of artificial general intelligence. All data, code and model checkpoints will be released upon acceptance.
ArXiv ID: 2510.15992 Authors: Ziming Dai, Tuo Zhang, Fei Gao, Xingyi Cai, Xiaofei Wang, Cheng Zhang, Wenyu Wang, Chengjie Zang
Abstract: arXiv:2510.15992v1 Announce Type: new Abstract: The growing industrial demand for customized and cost-efficient large language models (LLMs) is fueled by the rise of vertical, domain-specific tasks and the need to optimize performance under constraints such as latency and budget. Knowledge distillation, as an efficient model compression and transfer technique, offers a feasible solution. However, existing distillation frameworks often require manual intervention and struggle to meet such complex user-defined distillation requirements. To bridge this gap, we propose Stratos, an end-to-end LLM distillation pipeline that automates server and model selection, knowledge distillation, and deployment in distributed cloud environments. Given user-defined constraints on model performance and system budget, Stratos automatically selects Pareto-optimal servers, dynamically matches teacher-student pairs, and adapts distillation strategies based on task complexity to optimize cloud hosting. Experiments show that Stratos produces a student model that achieves four times the accuracy of its GPT-4o teacher baseline on a rare, domain-specific Mahjong reasoning task with reverse synthetic data and knowledge injection. Moreover, it achieves reduced latency and cost without compromising accuracy. These results highlight its promise for vertical-domain LLM deployment.
ArXiv ID: 2510.16824 Authors: Yingxu Wang, Kunyu Zhang, Jiaxin Huang, Nan Yin, Siwei Liu, Eran Segal
Abstract: arXiv:2510.16824v1 Announce Type: new Abstract: Multimodal molecular representation learning, which jointly models molecular graphs and their textual descriptions, enhances predictive accuracy and interpretability by enabling more robust and reliable predictions of drug toxicity, bioactivity, and physicochemical properties through the integration of structural and semantic information. However, existing multimodal methods suffer from two key limitations: (1) they typically perform cross-modal interaction only at the final encoder layer, thus overlooking hierarchical semantic dependencies; (2) they lack a unified prototype space for robust alignment between modalities. To address these limitations, we propose ProtoMol, a prototype-guided multimodal framework that enables fine-grained integration and consistent semantic alignment between molecular graphs and textual descriptions. ProtoMol incorporates dual-branch hierarchical encoders, utilizing Graph Neural Networks to process structured molecular graphs and Transformers to encode unstructured texts, resulting in comprehensive layer-wise representations. Then, ProtoMol introduces a layer-wise bidirectional cross-modal attention mechanism that progressively aligns semantic features across layers. Furthermore, a shared prototype space with learnable, class-specific anchors is constructed to guide both modalities toward coherent and discriminative representations. Extensive experiments on multiple benchmark datasets demonstrate that ProtoMol consistently outperforms state-of-the-art baselines across a variety of molecular property prediction tasks.
ArXiv ID: 2510.16774 Authors: Yuguang Yue, Irakli Salia, Samuel Hunt, Christopher Green, Wenzhe Shi, Jonathan J Hunt
Abstract: arXiv:2510.16774v1 Announce Type: new Abstract: We argue that 3-D first-person video games are a challenging environment for real-time multi-modal reasoning. We first describe our dataset of human game-play, collected across a large variety of 3-D first-person games, which is both substantially larger and more diverse compared to prior publicly disclosed datasets, and contains text instructions. We demonstrate that we can learn an inverse dynamics model from this dataset, which allows us to impute actions on a much larger dataset of publicly available videos of human game play that lack recorded actions. We then train a text-conditioned agent for game playing using behavior cloning, with a custom architecture capable of realtime inference on a consumer GPU. We show the resulting model is capable of playing a variety of 3-D games and responding to text input. Finally, we outline some of the remaining challenges such as long-horizon tasks and quantitative evaluation across a large set of games.
ArXiv ID: 2510.15948 Authors: MingSheng Li, Guangze Zhao, Sichen Liu
Abstract: arXiv:2510.15948v1 Announce Type: new Abstract: Large Vision-Language Models (LVLMs) have achieved remarkable progress in multimodal perception and generation, yet their safety alignment remains a critical challenge.Existing defenses and vulnerable to multimodal jailbreaks, as visual inputs introduce new attack surfaces, reasoning chains lack safety supervision, and alignment often degrades under modality fusion.To overcome these limitation, we propose VisuoAlign, a framework for multi-modal safety alignment via prompt-guided tree search.VisuoAlign embeds safety constrains into the reasoning process through visual-textual interactive prompts, employs Monte Carlo Tree Search(MCTS) to systematically construct diverse safety-critical prompt trajectories, and introduces prompt-based scaling to ensure real-time risk detection and compliant responses.Extensive experiments demonstrate that VisuoAlign proactively exposes risks, enables comprehensive dataset generation, and significantly improves the robustness of LVLMs against complex cross-modal threats.
ArXiv ID: 2510.17191 Authors: Peiru Zheng, Yun Zhao, Zhan Gong, Hong Zhu, Shaohua Wu
Abstract: arXiv:2510.17191v1 Announce Type: new Abstract: End-to-end autonomous driving has emerged as a promising paradigm for achieving robust and intelligent driving policies. However, existing end-to-end methods still face significant challenges, such as suboptimal decision-making in complex scenarios. In this paper,we propose SimpleVSF (Simple VLM-Scoring Fusion), a novel framework that enhances end-to-end planning by leveraging the cognitive capabilities of Vision-Language Models (VLMs) and advanced trajectory fusion techniques. We utilize the conventional scorers and the novel VLM-enhanced scorers. And we leverage a robust weight fusioner for quantitative aggregation and a powerful VLM-based fusioner for qualitative, context-aware decision-making. As the leading approach in the ICCV 2025 NAVSIM v2 End-to-End Driving Challenge, our SimpleVSF framework demonstrates state-of-the-art performance, achieving a superior balance between safety, comfort, and efficiency.
ArXiv ID: 2510.16039 Authors: Xiangyuan Peng, Xingsi Dong, Si Wu
Abstract: arXiv:2510.16039v1 Announce Type: new Abstract: We propose Grid-like Code Quantization (GCQ), a brain-inspired method for compressing observation-action sequences into discrete representations using grid-like patterns in attractor dynamics. Unlike conventional vector quantization approaches that operate on static inputs, GCQ performs spatiotemporal compression through an action-conditioned codebook, where codewords are derived from continuous attractor neural networks and dynamically selected based on actions. This enables GCQ to jointly compress space and time, serving as a unified world model. The resulting representation supports long-horizon prediction, goal-directed planning, and inverse modeling. Experiments across diverse tasks demonstrate GCQ's effectiveness in compact encoding and downstream performance. Our work offers both a computational tool for efficient sequence modeling and a theoretical perspective on the formation of grid-like codes in neural systems.
ArXiv ID: 2510.16617 Authors: Ruihan Zhao, Tyler Ingebrand, Sandeep Chinchali, Ufuk Topcu
Abstract: arXiv:2510.16617v1 Announce Type: new Abstract: Vision-Language-Action (VLA) models trained on large robot datasets promise general-purpose, robust control across diverse domains and embodiments. However, existing approaches often fail out-of-the-box when deployed in novel environments, embodiments, or tasks. We introduce Mixture of Skills VLA (MoS-VLA), a framework that represents robot manipulation policies as linear combinations of a finite set of learned basis functions. During pretraining, MoS-VLA jointly learns these basis functions across datasets from the Open X-Embodiment project, producing a structured skill space. At test time, adapting to a new task requires only a single expert demonstration. The corresponding skill representation is then inferred via a lightweight convex optimization problem that minimizes the L1 action error, without requiring gradient updates. This gradient-free adaptation incurs minimal overhead while enabling rapid instantiation of new skills. Empirically, MoS-VLA achieves lower action-prediction error on five out of five unseen datasets and succeeds in both simulation and real-robot tasks where a pretrained VLA model fails outright. Project page: mos-vla.github.io/
ArXiv ID: 2510.16077 Authors: Naeem Paeedeh, Mahardhika Pratama, Weiping Ding, Jimmy Cao, Wolfgang Mayer, Ryszard Kowalczyk
Abstract: arXiv:2510.16077v1 Announce Type: new Abstract: Domain Incremental Learning (DIL) is a continual learning sub-branch that aims to address never-ending arrivals of new domains without catastrophic forgetting problems. Despite the advent of parameter-efficient fine-tuning (PEFT) approaches, existing works create task-specific LoRAs overlooking shared knowledge across tasks. Inaccurate selection of task-specific LORAs during inference results in significant drops in accuracy, while existing works rely on linear or prototype-based classifiers, which have suboptimal generalization powers. Our paper proposes continual knowledge consolidation low rank adaptation (CONEC-LoRA) addressing the DIL problems. CONEC-LoRA is developed from consolidations between task-shared LORA to extract common knowledge and task-specific LORA to embrace domain-specific knowledge. Unlike existing approaches, CONEC-LoRA integrates the concept of a stochastic classifier whose parameters are sampled from a distribution, thus enhancing the likelihood of correct classifications. Last but not least, an auxiliary network is deployed to optimally predict the task-specific LoRAs for inferences and implements the concept of a different-depth network structure in which every layer is connected with a local classifier to take advantage of intermediate representations. This module integrates the ball-generator loss and transformation module to address the synthetic sample bias problem. Our rigorous experiments demonstrate the advantage of CONEC-LoRA over prior arts in 4 popular benchmark problems with over 5% margins.
ArXiv ID: 2510.17002 Authors: Chang Liu, Danial Chitnis
Abstract: arXiv:2510.17002v1 Announce Type: new Abstract: Circuit schematics play a crucial role in analog integrated circuit design, serving as the primary medium for human understanding and verification of circuit functionality. While recent large language model (LLM)-based approaches have shown promise in circuit topology generation and device sizing, most rely solely on textual representations such as SPICE netlists, which lack visual interpretability for circuit designers. To address this limitation, we propose EEschematic, an AI agent for automatic analog schematic generation based on a Multimodal Large Language Model (MLLM). EEschematic integrates textual, visual, and symbolic modalities to translate SPICE netlists into schematic diagrams represented in a human-editable format. The framework uses six analog substructure examples for few-shot placement and a Visual Chain-of-Thought (VCoT) strategy to iteratively refine placement and wiring, enhancing schematic clarity and symmetry. Experimental results on representative analog circuits, including a CMOS inverter, a five-transistor operational transconductance amplifier (5T-OTA), and a telescopic cascode amplifier, demonstrate that EEschematic produces schematics with high visual quality and structural correctness.
ArXiv ID: 2510.16990 Authors: Xuying Ning, Dongqi Fu, Tianxin Wei, Wujiang Xu, Jingrui He
Abstract: arXiv:2510.16990v1 Announce Type: new Abstract: Real-world multimodal data usually exhibit complex structural relationships beyond traditional one-to-one mappings like image-caption pairs. Entities across modalities interact in intricate ways, with images and text forming diverse interconnections through contextual dependencies and co-references. Graphs provide powerful structural information for modeling intra-modal and inter-modal relationships. However, previous works fail to distinguish multi-hop neighbors and treat the graph as a standalone modality, which fragments the overall understanding. This limitation presents two key challenges in multimodal learning: (1) integrating structural information from multi-hop neighbors into foundational models, and (2) fusing modality-specific information in a principled manner. To address these challenges, we revisit the role of graphs in multimodal learning within the era of foundation models and propose Graph4MM, a graph-based multimodal learning framework. To be specific, we introduce Hop-Diffused Attention, which integrates multi-hop structural information into self-attention through causal masking and hop diffusion. Furthermore, we design MM-QFormer, a multi-mapping querying transformer for cross-modal fusion. Through theoretical and empirical analysis, we show that leveraging structures to integrate both intra- and inter-modal interactions improves multimodal understanding beyond treating them as a standalone modality. Experiments on both generative and discriminative tasks show that Graph4MM outperforms larger VLMs, LLMs, and multimodal graph baselines, achieving a 6.93% average improvement.
ArXiv ID: 2510.15964 Authors: Tuowei Wang, Kun Li, Zixu Hao, Donglin Bai, Ju Ren, Yaoxue Zhang, Ting Cao, Mao Yang
Abstract: arXiv:2510.15964v1 Announce Type: new Abstract: The adaptation of pre-trained large language models (LLMs) to diverse downstream tasks via fine-tuning is critical for numerous applications. However, the inefficiency of parameter-efficient fine-tuning (PEFT) techniques presents significant challenges in terms of time investments and operational costs. In this paper, we first introduce a nuanced form of sparsity, termed Shadowy Sparsity, which is distinctive in fine-tuning and has not been adequately addressed for acceleration. Under Shadowy Sparsity, we propose Long Exposure, an efficient system to accelerate PEFT for LLMs. Long Exposure comprises three key components: Shadowy-sparsity Exposer employs a prolonged sensing range to capture more sparsity details under shadowy sparsity; Sequence-oriented Predictor provides efficient yet accurate predictions to handle large sequence inputs and constantly-evolving parameters; and Dynamic-aware Operator facilitates more structured computational patterns and coalesced memory accesses, addressing dynamic sparse operations. Extensive evaluations show that Long Exposure outperforms state-of-the-arts with up to a $2.49\times$ speedup in end-to-end fine-tuning, offering promising advancements in accelerating PEFT for LLMs.
ArXiv ID: 2510.17189 Authors: Wenxun Wang, Shuchang Zhou, Wenyu Sun, Peiqin Sun, Yongpan Liu
Abstract: arXiv:2510.17189v1 Announce Type: new Abstract: Transformers have shown remarkable performance in both natural language processing (NLP) and computer vision (CV) tasks. However, their real-time inference speed and efficiency are limited due to the inefficiency in Softmax and Layer Normalization (LayerNorm). Previous works based on function approximation suffer from inefficient implementation as they place emphasis on computation while disregarding memory overhead concerns. Moreover, such methods rely on retraining to compensate for approximation error which can be costly and inconvenient. In this paper, we present SOLE, a hardware-software co-design for Softmax and LayerNorm which is composed of E2Softmax and AILayerNorm. E2Softmax utilizes log2 quantization of exponent function and log-based division to approximate Softmax while AILayerNorm adopts low-precision statistic calculation. Compared with state-of-the-art designs, we achieve both low-precision calculation and low bit-width storage on Softmax and LayerNorm. Experiments show that SOLE maintains inference accuracy without retraining while offering orders of magnitude speedup and energy savings over GPU, achieving 3.04x, 3.86x energy-efficiency improvements and 2.82x, 3.32x area-efficiency improvements over prior state-of-the-art custom hardware for Softmax and LayerNorm, respectively.
ArXiv ID: 2510.16045 Authors: Mengtao Lv, Ruiqi Zhu, Xinyu Wang, Yun Li
Abstract: arXiv:2510.16045v1 Announce Type: new Abstract: Large language models (LLMs) have demonstrated remarkable capabilities in various kinds of tasks, while the billion or even trillion parameters bring storage and efficiency bottlenecks for inference. Quantization, particularly floating-point quantization, is known to be capable of speeding up LLM inference by reducing memory footprint and data movement during the inference process. For the first time, we advance the floating-point quantization exploration from integer bitwidths to non-integer bit-widths, namely AMS-Quant, to further approach the quantization sweet spot. AMS-Quant incorporates two novel techniques to put it into effect: (1) it proposes Mantissa-bit Sharing, which groups k quantized weights and lets them share the least significant mantissa bit, allowing us to further approach the minimum quantization bit-width without accuracy loss. (2) It introduces Adaptive Searching, which employs an offline optimization strategy to minimize the accuracy degradation introduced by sharing. Moreover, AMS-Quant is also prototyped as efficient CUDA Linear kernels, which translates memory savings into wall-clock latency reduction by reducing memory access. Extensive experiments on large-scale datasets and models show that AMS-Quant can quantize the model to FP-5.33-e2m3 and FP4.25-e2m2, and significantly speed up the LLM decoding over FP16 inference (2.8x and 3.2x), with negligible accuracy loss.
ArXiv ID: 2510.17150 Authors: Heng Zhang, Wei-Hsing Huang, Gokhan Solak, Arash Ajoudani
Abstract: arXiv:2510.17150v1 Announce Type: new Abstract: We present OmniVIC, a universal variable impedance controller (VIC) enhanced by a vision language model (VLM), which improves safety and adaptation in any contact-rich robotic manipulation task to enhance safe physical interaction. Traditional VIC have shown advantages when the robot physically interacts with the environment, but lack generalization in unseen, complex, and unstructured safe interactions in universal task scenarios involving contact or uncertainty. To this end, the proposed OmniVIC interprets task context derived reasoning from images and natural language and generates adaptive impedance parameters for a VIC controller. Specifically, the core of OmniVIC is a self-improving Retrieval-Augmented Generation(RAG) and in-context learning (ICL), where RAG retrieves relevant prior experiences from a structured memory bank to inform the controller about similar past tasks, and ICL leverages these retrieved examples and the prompt of current task to query the VLM for generating context-aware and adaptive impedance parameters for the current manipulation scenario. Therefore, a self-improved RAG and ICL guarantee OmniVIC works in universal task scenarios. The impedance parameter regulation is further informed by real-time force/torque feedback to ensure interaction forces remain within safe thresholds. We demonstrate that our method outperforms baselines on a suite of complex contact-rich tasks, both in simulation and on real-world robotic tasks, with improved success rates and reduced force violations. OmniVIC takes a step towards bridging high-level semantic reasoning and low-level compliant control, enabling safer and more generalizable manipulation. Overall, the average success rate increases from 27% (baseline) to 61.4% (OmniVIC).
ArXiv ID: 2510.16092 Authors: Devvrit Khatri, Pranamya Kulkarni, Nilesh Gupta, Yerram Varun, Liqian Peng, Jay Yagnik, Praneeth Netrapalli, Cho-Jui Hsieh, Alec Go, Inderjit S Dhillon, Aditya Kusupati, Prateek Jain
Abstract: arXiv:2510.16092v1 Announce Type: new Abstract: Large Language Models (LLMs) have been shown to be able to learn different tasks without explicit finetuning when given many input-output examples / demonstrations through In-Context Learning (ICL). Increasing the number of examples, called ``shots'', improves downstream task performance but incurs higher memory and computational costs. In this work, we study an approach to improve the memory and computational efficiency of ICL inference by compressing the many-shot prompts. Given many shots comprising t tokens, our goal is to generate a m soft-token summary, where m < t. We first show that existing prompt compression methods are ineffective for many-shot compression, and simply using fewer shots as a baseline is surprisingly strong. To achieve effective compression, we find that: (a) a stronger compressor model with more trainable parameters is necessary, and (b) compressing many-shot representations at each transformer layer enables more fine-grained compression by providing each layer with its own compressed representation. Based on these insights, we propose MemCom, a layer-wise compression method. We systematically evaluate various compressor models and training approaches across different model sizes (2B and 7B), architectures (Gemma and Mistral), many-shot sequence lengths (3k-6k tokens), and compression ratios (3x to 8x). MemCom outperforms strong baselines across all compression ratios on multiple classification tasks with large label sets. Notably, while baseline performance degrades sharply at higher compression ratios, often by over 20-30%, MemCom maintains high accuracy with minimal degradation, typically dropping by less than 10%.
ArXiv ID: 2510.17590 Authors: Mir Nafis Sharear Shopnil, Sharad Duwal, Abhishek Tyagi, Adiba Mahbub Proma
Abstract: arXiv:2510.17590v1 Announce Type: new Abstract: Misinformation spreads across web platforms through billions of daily multimodal posts that combine text and images, overwhelming manual fact-checking capacity. Supervised detection models require domain-specific training data and fail to generalize across diverse manipulation tactics. We present MIRAGE, an inference-time, model-pluggable agentic framework that decomposes multimodal verification into four sequential modules: visual veracity assessment detects AI-generated images, cross-modal consistency analysis identifies out-of-context repurposing, retrieval-augmented factual checking grounds claims in web evidence through iterative question generation, and a calibrated judgment module integrates all signals. MIRAGE orchestrates vision-language model reasoning with targeted web retrieval, outputs structured and citation-linked rationales. On MMFakeBench validation set (1,000 samples), MIRAGE with GPT-4o-mini achieves 81.65% F1 and 75.1% accuracy, outperforming the strongest zero-shot baseline (GPT-4V with MMD-Agent at 74.0% F1) by 7.65 points while maintaining 34.3% false positive rate versus 97.3% for a judge-only baseline. Test set results (5,000 samples) confirm generalization with 81.44% F1 and 75.08% accuracy. Ablation studies show visual verification contributes 5.18 F1 points and retrieval-augmented reasoning contributes 2.97 points. Our results demonstrate that decomposed agentic reasoning with web retrieval can match supervised detector performance without domain-specific training, enabling misinformation detection across modalities where labeled data remains scarce.
ArXiv ID: 2510.16769 Authors: Shuo Han, Yukun Cao, Zezhong Ding, Zengyi Gao, S Kevin Zhou, Xike Xie
Abstract: arXiv:2510.16769v1 Announce Type: new Abstract: Vision-language models (VLMs) have shown promise in graph understanding, but remain limited by input-token constraints, facing scalability bottlenecks and lacking effective mechanisms to coordinate textual and visual modalities. To address these challenges, we propose GraphVista, a unified framework that enhances both scalability and modality coordination in graph understanding. For scalability, GraphVista organizes graph information hierarchically into a lightweight GraphRAG base, which retrieves only task-relevant textual descriptions and high-resolution visual subgraphs, compressing redundant context while preserving key reasoning elements. For modality coordination, GraphVista introduces a planning agent that routes tasks to the most suitable modality-using the text modality for simple property reasoning and the visual modality for local and structurally complex reasoning grounded in explicit topology. Extensive experiments demonstrate that GraphVista scales to large graphs, up to $200\times$ larger than those used in existing benchmarks, and consistently outperforms existing textual, visual, and fusion-based methods, achieving up to $4.4\times$ quality improvement over the state-of-the-art baselines by fully exploiting the complementary strengths of both modalities.
ArXiv ID: 2510.15945 Authors: Guangya Wan, Zixin Stephen Xu, Sasa Zorc, Manel Baucells, Mengxuan Hu, Hao Wang, Sheng Li
Abstract: arXiv:2510.15945v1 Announce Type: new Abstract: Sampling multiple responses is a common way to improve LLM output quality, but it comes at the cost of additional computation. The key challenge is deciding when to stop generating new samples to balance accuracy gains against efficiency. To address this, we introduce BEACON (Bayesian Efficient Adaptive Criterion for Optimal N-stopping), a principled adaptive sampling framework grounded in Sequential Search with Bayesian Learning. BEACON sequentially generates responses from the policy LLM, updates posterior belief over reward distributions in real time without further training, and determines when to stop by weighing expected gains against computational cost. Sampling terminates once the marginal utility of further exploration no longer justifies the expense. We establish both theoretical optimality guarantees and practical tractability, and show empirically that BEACON reduces average sampling by up to 80% while maintaining response quality. We further demonstrate BEACON's utility for cost-efficient preference data generation and outline practical extensions, offering actionable insights for future researchers.
ArXiv ID: 2510.16089 Authors: William Hoy, Nurcin Celik
Abstract: arXiv:2510.16089v1 Announce Type: new Abstract: Large language models (LLMs) increasingly require mechanisms for continual adaptation without full retraining. However, sequential updates can lead to catastrophic forgetting, where new edits degrade previously acquired knowledge. This work presents STABLE, a gated continual self editing framework that constrains forgetting during sequential updates using parameter efficient fine tuning via Low Rank Adaptation (LoRA; see arXiv:2106.09685). Each candidate edit is evaluated against a stability budget using one of three metrics: (i) Exact Match (EM) drop, capturing factual accuracy loss; (ii) bits increase, reflecting reduced model confidence; and (iii) KL divergence, quantifying distributional drift between the base and adapted models. If a threshold is exceeded, the LoRA update is rescaled through a clipping procedure or rejected. Experiments on the Qwen-2.5-7B model show that gating effectively mitigates forgetting while preserving adaptability. EM based gating achieved the highest cumulative performance in short continual learning sequences. Our results show that different gating strategies can achieve comparable distribution shift (measured by KL divergence) while producing different accuracy outcomes, highlighting the importance of gating design in continual adaptation. This approach offers a principled method for continual model editing, enabling LLMs to integrate new knowledge while maintaining reliability. Code: https://github.com/Bhoy1/STABLE
ArXiv ID: 2510.16753 Authors: Wei Huang, Peining Li, Meiyu Liang, Xu Hou, Junping Du, Yingxia Shao, Guanhua Ye, Wu Liu, Kangkang Lu, Yang Yu
Abstract: arXiv:2510.16753v1 Announce Type: new Abstract: Multimodal Knowledge Graphs (MKGs) extend traditional knowledge graphs by incorporating visual and textual modalities, enabling richer and more expressive entity representations. However, existing MKGs often suffer from incompleteness, which hinder their effectiveness in downstream tasks. Therefore, multimodal knowledge graph completion (MKGC) task is receiving increasing attention. While large language models (LLMs) have shown promise for knowledge graph completion (KGC), their application to the multimodal setting remains underexplored. Moreover, applying Multimodal Large Language Models (MLLMs) to the task of MKGC introduces significant challenges: (1) the large number of image tokens per entity leads to semantic noise and modality conflicts, and (2) the high computational cost of processing large token inputs. To address these issues, we propose Efficient Lightweight Multimodal Large Language Models (ELMM) for MKGC. ELMM proposes a Multi-view Visual Token Compressor (MVTC) based on multi-head attention mechanism, which adaptively compresses image tokens from both textual and visual views, thereby effectively reducing redundancy while retaining necessary information and avoiding modality conflicts. Additionally, we design an attention pruning strategy to remove redundant attention layers from MLLMs, thereby significantly reducing the inference cost. We further introduce a linear projection to compensate for the performance degradation caused by pruning. Extensive experiments on benchmark FB15k-237-IMG and WN18-IMG demonstrate that ELMM achieves state-of-the-art performance while substantially improving computational efficiency, establishing a new paradigm for multimodal knowledge graph completion.
ArXiv ID: 2510.16907 Authors: Kangrui Wang, Pingyue Zhang, Zihan Wang, Yaning Gao, Linjie Li, Qineng Wang, Hanyang Chen, Chi Wan, Yiping Lu, Zhengyuan Yang, Lijuan Wang, Ranjay Krishna, Jiajun Wu, Li Fei-Fei, Yejin Choi, Manling Li
Abstract: arXiv:2510.16907v1 Announce Type: new Abstract: A key challenge in training Vision-Language Model (VLM) agents, compared to Language Model (LLM) agents, lies in the shift from textual states to complex visual observations. This transition introduces partial observability and demands robust world modeling. We ask: Can VLM agents construct internal world models through explicit visual state reasoning? To address this question, we architecturally enforce and reward the agent's reasoning process via reinforcement learning (RL), formulating it as a Partially Observable Markov Decision Process (POMDP). We find that decomposing the agent's reasoning into State Estimation ("what is the current state?") and Transition Modeling ("what comes next?") is critical for success, as demonstrated through five reasoning strategies. Our investigation into how agents represent internal beliefs reveals that the optimal representation is task-dependent: Natural Language excels at capturing semantic relationships in general tasks, while Structured formats are indispensable for precise manipulation and control. Building on these insights, we design a World Modeling Reward that provides dense, turn-level supervision for accurate state prediction, and introduce Bi-Level General Advantage Estimation (Bi-Level GAE) for turn-aware credit assignment. Through this form of visual state reasoning, a 3B-parameter model achieves a score of 0.82 across five diverse agent benchmarks, representing a 3$\times$ improvement over its untrained counterpart (0.21) and outperforming proprietary reasoning models such as GPT-5 (0.75), Gemini 2.5 Pro (0.67) and Claude 4.5 (0.62). All experiments are conducted within our VAGEN framework, a scalable system for training and analyzing multi-turn VLM agents in diverse visual environments. Code and data are publicly available at https://vagen-ai.github.io.
ArXiv ID: 2510.15962 Authors: Zhuxuanzi Wang, Mingqiao Mo, Xi Xiao, Chen Liu, Chenrui Ma, Yunbei Zhang, Xiao Wang, Smita Krishnaswamy, Tianyang Wang
Abstract: arXiv:2510.15962v1 Announce Type: new Abstract: Parameter-efficient fine-tuning (PEFT) has become the standard approach for adapting large language models under limited compute and memory budgets. Although previous methods improve efficiency through low-rank updates, quantization, or heuristic budget reallocation, they often decouple the allocation of capacity from the way updates evolve during training. In this work, we introduce CTR-LoRA, a framework guided by curvature trust region that integrates rank scheduling with stability-aware optimization. CTR-LoRA allocates parameters based on marginal utility derived from lightweight second-order proxies and constrains updates using a Fisher/Hessian-metric trust region. Experiments on multiple open-source backbones (7B-13B), evaluated on both in-distribution and out-of-distribution benchmarks, show consistent improvements over strong PEFT baselines. In addition to increased accuracy, CTR-LoRA enhances training stability, reduces memory requirements, and achieves higher throughput, positioning it on the Pareto frontier of performance and efficiency. These results highlight a principled path toward more robust and deployable PEFT.
ArXiv ID: 2510.16448 Authors: Yongxiang Hua, Haoyu Cao, Zhou Tao, Bocheng Li, Zihao Wu, Chaohu Liu, Linli Xu
Abstract: arXiv:2510.16448v1 Announce Type: new Abstract: Sparse Mixture of Experts (sMoE) has become a pivotal approach for scaling large vision-language models, offering substantial capacity while maintaining computational efficiency through dynamic, sparse activation of experts. However, existing routing mechanisms, typically based on similarity scoring, struggle to effectively capture the underlying input structure. This limitation leads to a trade-off between expert specialization and balanced computation, hindering both scalability and performance. We propose Input Domain Aware MoE, a novel routing framework that leverages a probabilistic mixture model to better partition the input space. By modeling routing probabilities as a mixture of distributions, our method enables experts to develop clear specialization boundaries while achieving balanced utilization. Unlike conventional approaches, our routing mechanism is trained independently of task-specific objectives, allowing for stable optimization and decisive expert assignments. Empirical results on vision-language tasks demonstrate that our method consistently outperforms existing sMoE approaches, achieving higher task performance and improved expert utilization balance.
ArXiv ID: 2510.16805 Authors: Mariam Rakka, Marios Fournarakis, Olga Krestinskaya, Jinane Bazzi, Khaled N. Salama, Fadi Kurdahi, Ahmed M. Eltawil, Mohammed E. Fouda
Abstract: arXiv:2510.16805v1 Announce Type: new Abstract: The rapid scaling of language models (LMs) has resulted in unprecedented computational, memory, and energy requirements, making their training and deployment increasingly unsustainable. Quantization has emerged as an essential compression technique to reduce model size, alleviate memory bottlenecks, and accelerate inference. However, while uniform low-bit quantization (e.g., INT8, INT4) provides significant efficiency gains, it can degrade accuracy in sensitive components of transformer-based LMs. Mixed-precision quantization offers a promising alternative by selectively allocating precision across layers or within tensors to balance efficiency and accuracy. This survey provides a comprehensive overview of Mixed-Precision quantization frameworks for LMs (MXPLMs). We first review quantization fundamentals, including uniform and non-uniform quantizers, quantization granularity, and methods widely used in post-training quantization. We then categorize and compare recent MXPLM frameworks according to their bit allocation strategies and precision configurations across weights, activations, and key-value caches. A comparative analysis highlights differences in perplexity, zero-shot task performance, and deployment trade-offs. Furthermore, we contrast MXPLMs with earlier mixed-precision quantization methods for deep neural networks, identifying strategies that transfer and those that face challenges in the LM setting. Finally, we summarize open issues and future directions, including hardware-aware design, activation quantization, and scalable optimization methods for billion-parameter models. By consolidating recent advances, this work serves as a reference for understanding the current landscape and research prospects of mixed-precision quantization for large-scale language models.
ArXiv ID: 2510.16411 Authors: Minh-Khoi Nguyen-Nhat, Rachel S. Y. Teo, Laziz Abdullaev, Maurice Mok, Viet-Hoang Tran, Tan Minh Nguyen
Abstract: arXiv:2510.16411v1 Announce Type: new Abstract: Sparse Mixture of Experts (SMoE) has emerged as a promising solution to achieving unparalleled scalability in deep learning by decoupling model parameter count from computational cost. By activating only a small subset of parameters per sample, SMoE enables significant growth in model capacity while maintaining efficiency. However, SMoE struggles to adapt to distributional shifts, leading to reduced robustness under data contamination. In this work, we introduce SymphonySMoE, a novel family of SMoE that introduces a social graph to model interactions among experts. This graph-based structure enhances the token routing process, addressing the robustness challenges that are inherent in conventional SMoE designs. SymphonySMoE is lightweight, modular, and integrates seamlessly with existing SMoE-based models such as the XMoE and the Generalist Language Model. We provide both theoretical analysis and empirical evidence demonstrating SymphonySMoE's advantages over baseline SMoE. Extensive experiments on language modeling and visual instruction tuning validate our method's effectiveness. We further highlight the scalability of SymphonySMoE to models with 4.2 and 7.4 billion parameters, showcasing its applicability in fine-tuning tasks for large-scale systems.
ArXiv ID: 2510.16281 Authors: Yilin Wu, Anqi Li, Tucker Hermans, Fabio Ramos, Andrea Bajcsy, Claudia P'erez-D'Arpino
Abstract: arXiv:2510.16281v1 Announce Type: new Abstract: Reasoning Vision Language Action (VLA) models improve robotic instruction-following by generating step-by-step textual plans before low-level actions, an approach inspired by Chain-of-Thought (CoT) reasoning in language models. Yet even with a correct textual plan, the generated actions can still miss the intended outcomes in the plan, especially in out-of-distribution (OOD) scenarios. We formalize this phenomenon as a lack of embodied CoT faithfulness, and introduce a training-free, runtime policy steering method for reasoning-action alignment. Given a reasoning VLA's intermediate textual plan, our framework samples multiple candidate action sequences from the same model, predicts their outcomes via simulation, and uses a pre-trained Vision-Language Model (VLM) to select the sequence whose outcome best aligns with the VLA's own textual plan. Only executing action sequences that align with the textual reasoning turns our base VLA's natural action diversity from a source of error into a strength, boosting robustness to semantic and visual OOD perturbations and enabling novel behavior composition without costly re-training. We also contribute a reasoning-annotated extension of LIBERO-100, environment variations tailored for OOD evaluation, and demonstrate up to 15% performance gain over prior work on behavior composition tasks and scales with compute and data diversity. Project Website at: https://yilin-wu98.github.io/steering-reasoning-vla/
ArXiv ID: 2510.17771 Authors: Zhining Liu, Ziyi Chen, Hui Liu, Chen Luo, Xianfeng Tang, Suhang Wang, Joy Zeng, Zhenwei Dai, Zhan Shi, Tianxin Wei, Benoit Dumoulin, Hanghang Tong
Abstract: arXiv:2510.17771v1 Announce Type: new Abstract: Vision-Language Models (VLMs) achieve strong results on multimodal tasks such as visual question answering, yet they can still fail even when the correct visual evidence is present. In this work, we systematically investigate whether these failures arise from not perceiving the evidence or from not leveraging it effectively. By examining layer-wise attention dynamics, we find that shallow layers focus primarily on text, while deeper layers sparsely but reliably attend to localized evidence regions. Surprisingly, VLMs often perceive the visual evidence when outputting incorrect answers, a phenomenon we term ``seeing but not believing'' that widely exists in major VLM families. Building on this, we introduce an inference-time intervention that highlights deep-layer evidence regions through selective attention-based masking. It requires no training and consistently improves accuracy across multiple families, including LLaVA, Qwen, Gemma, and InternVL. These results show that VLMs encode reliable evidence internally but under-utilize it, making such signals explicit can bridge the gap between perception and reasoning, advancing the diagnostic understanding and reliability of VLMs.
ArXiv ID: 2510.16138 Authors: Dung V. Nguyen, Anh T. Nguyen, Minh H. Nguyen, Luc Q. Nguyen, Shiqi Jiang, Ethan Fetaya, Linh Duy Tran, Gal Chechik, Tan M. Nguyen
Abstract: arXiv:2510.16138v1 Announce Type: new Abstract: Existing expert merging strategies for Sparse Mixture of Experts (SMoE) typically rely on input-dependent or input-independent averaging of expert parameters, but often lack a principled weighting mechanism. In this work, we reinterpret expert merging through the lens of game theory, revealing cooperative and competitive dynamics among experts. Based on this perspective, we introduce Nash Merging of Experts (NAMEx), a novel framework that incorporates Nash Bargaining into the merging process, enabling more balanced and efficient collaboration among experts. Additionally, we incorporate complex momentum into NAMEx to accelerate expert propagation with theoretical guarantees for convergence. Extensive experiments across language modelling, text classification, image classification, and zero-shot robustness under data corruption show that NAMEx consistently outperforms competing methods while integrating seamlessly with popular MoE architectures. Finally, we demonstrate NAMEx's scalability by applying it to large-scale systems, including Qwen1.5-MoE (14B) and DeepSeek-MoE (16B), where it proves effective in both zero-shot and fine-tuning settings.
ArXiv ID: 2510.17439 Authors: Zhengshen Zhang, Hao Li, Yalun Dai, Zhengbang Zhu, Lei Zhou, Chenchen Liu, Dong Wang, Francis E. H. Tay, Sijin Chen, Ziwei Liu, Yuxiao Liu, Xinghang Li, Pan Zhou
Abstract: arXiv:2510.17439v1 Announce Type: new Abstract: Existing vision-language-action (VLA) models act in 3D real-world but are typically built on 2D encoders, leaving a spatial reasoning gap that limits generalization and adaptability. Recent 3D integration techniques for VLAs either require specialized sensors and transfer poorly across modalities, or inject weak cues that lack geometry and degrade vision-language alignment. In this work, we introduce FALCON (From Spatial to Action), a novel paradigm that injects rich 3D spatial tokens into the action head. FALCON leverages spatial foundation models to deliver strong geometric priors from RGB alone, and includes an Embodied Spatial Model that can optionally fuse depth, or pose for higher fidelity when available, without retraining or architectural changes. To preserve language reasoning, spatial tokens are consumed by a Spatial-Enhanced Action Head rather than being concatenated into the vision-language backbone. These designs enable FALCON to address limitations in spatial representation, modality transferability, and alignment. In comprehensive evaluations across three simulation benchmarks and eleven real-world tasks, our proposed FALCON achieves state-of-the-art performance, consistently surpasses competitive baselines, and remains robust under clutter, spatial-prompt conditioning, and variations in object scale and height.
ArXiv ID: 2510.17058 Authors: Hassan Hamad, Yuou Qiu, Peter A. Beerel, Keith M. Chugg
Abstract: arXiv:2510.17058v1 Announce Type: new Abstract: While advancements in quantization have significantly reduced the computational costs of inference in deep learning, training still predominantly relies on complex floating-point arithmetic. Low-precision fixed-point training presents a compelling alternative. This work introduces a novel enhancement in low-precision logarithmic fixed-point training, geared towards future hardware accelerator designs. We propose incorporating bitwidth in the design of approximations to arithmetic operations. To this end, we introduce a new hardware-friendly, piece-wise linear approximation for logarithmic addition. Using simulated annealing, we optimize this approximation at different precision levels. A C++ bit-true simulation demonstrates training of VGG-11 and VGG-16 models on CIFAR-100 and TinyImageNet, respectively, using 12-bit integer arithmetic with minimal accuracy degradation compared to 32-bit floating-point training. Our hardware study reveals up to 32.5% reduction in area and 53.5% reduction in energy consumption for the proposed LNS multiply-accumulate units compared to that of linear fixed-point equivalents.
ArXiv ID: 2510.17369 Authors: Haochen Su, Cristian Meo, Francesco Stella, Andrea Peirone, Kai Junge, Josie Hughes
Abstract: arXiv:2510.17369v1 Announce Type: new Abstract: Robotic systems are increasingly expected to operate in human-centered, unstructured environments where safety, adaptability, and generalization are essential. Vision-Language-Action (VLA) models have been proposed as a language guided generalized control framework for real robots. However, their deployment has been limited to conventional serial link manipulators. Coupled by their rigidity and unpredictability of learning based control, the ability to safely interact with the environment is missing yet critical. In this work, we present the deployment of a VLA model on a soft continuum manipulator to demonstrate autonomous safe human-robot interaction. We present a structured finetuning and deployment pipeline evaluating two state-of-the-art VLA models (OpenVLA-OFT and $\pi_0$) across representative manipulation tasks, and show while out-of-the-box policies fail due to embodiment mismatch, through targeted finetuning the soft robot performs equally to the rigid counterpart. Our findings highlight the necessity of finetuning for bridging embodiment gaps, and demonstrate that coupling VLA models with soft robots enables safe and flexible embodied AI in human-shared environments.
ArXiv ID: 2510.17801 Authors: Yulin Luo, Chun-Kai Fan, Menghang Dong, Jiayu Shi, Mengdi Zhao, Bo-Wen Zhang, Cheng Chi, Jiaming Liu, Gaole Dai, Rongyu Zhang, Ruichuan An, Kun Wu, Zhengping Che, Shaoxuan Xie, Guocai Yao, Zhongxia Zhao, Pengwei Wang, Guang Liu, Zhongyuan Wang, Tiejun Huang, Shanghang Zhang
Abstract: arXiv:2510.17801v1 Announce Type: new Abstract: Building robots that can perceive, reason, and act in dynamic, unstructured environments remains a core challenge. Recent embodied systems often adopt a dual-system paradigm, where System 2 handles high-level reasoning while System 1 executes low-level control. In this work, we refer to System 2 as the embodied brain, emphasizing its role as the cognitive core for reasoning and decision-making in manipulation tasks. Given this role, systematic evaluation of the embodied brain is essential. Yet existing benchmarks emphasize execution success, or when targeting high-level reasoning, suffer from incomplete dimensions and limited task realism, offering only a partial picture of cognitive capability. To bridge this gap, we introduce RoboBench, a benchmark that systematically evaluates multimodal large language models (MLLMs) as embodied brains. Motivated by the critical roles across the full manipulation pipeline, RoboBench defines five dimensions-instruction comprehension, perception reasoning, generalized planning, affordance prediction, and failure analysis-spanning 14 capabilities, 25 tasks, and 6092 QA pairs. To ensure realism, we curate datasets across diverse embodiments, attribute-rich objects, and multi-view scenes, drawing from large-scale real robotic data. For planning, RoboBench introduces an evaluation framework, MLLM-as-world-simulator. It evaluate embodied feasibility by simulating whether predicted plans can achieve critical object-state changes. Experiments on 14 MLLMs reveal fundamental limitations: difficulties with implicit instruction comprehension, spatiotemporal reasoning, cross-scenario planning, fine-grained affordance understanding, and execution failure diagnosis. RoboBench provides a comprehensive scaffold to quantify high-level cognition, and guide the development of next-generation embodied MLLMs. The project page is in https://robo-bench.github.io.
ArXiv ID: 2510.17802 Authors: Rui Pan, Yang Luo, Yuxing Liu, Yang You, Tong Zhang
Abstract: arXiv:2510.17802v1 Announce Type: new Abstract: Memory-efficient optimization is critical for training increasingly large language models (LLMs). A popular strategy involves gradient low-rank projection, storing only the projected optimizer states, with GaLore being a representative example. However, a significant drawback of many such methods is their lack of convergence guarantees, as various low-rank projection approaches introduce inherent biases relative to the original optimization algorithms, which contribute to performance gaps compared to full-parameter training. Aiming to tackle this problem, this paper investigates the layerwise sampling technique for debiasing low-rank projection mechanisms. In particular, an instantiation of the paradigm gives rise to a novel and unbiased low-rank optimization method built upon GaLore's mechanism and the Muon algorithm, named GaLore Unbiased with Muon (GUM). We theoretically prove our method matches the convergence guarantees of the base Muon algorithm while preserving the memory efficiency of low-rank techniques. Empirical experiments on LLM fine-tuning and pretraining also demonstrate non-trivial improvements over GaLore and even better performance than full-parameter training. Further investigation shows that the improvement of this technique comes from a more uniform distribution of knowledge inside layers, leading to more efficient utilization of the model parameter space and better memorization.
ArXiv ID: 2510.17394 Authors: Alejandro Guerra-Manzanares, Farah E. Shamout
Abstract: arXiv:2510.17394v1 Announce Type: new Abstract: The aim of multimodal neural networks is to combine diverse data sources, referred to as modalities, to achieve enhanced performance compared to relying on a single modality. However, training of multimodal networks is typically hindered by modality overfitting, where the network relies excessively on one of the available modalities. This often yields sub-optimal performance, hindering the potential of multimodal learning and resulting in marginal improvements relative to unimodal models. In this work, we present the Modality-Informed Learning ratE Scheduler (MILES) for training multimodal joint fusion models in a balanced manner. MILES leverages the differences in modality-wise conditional utilization rates during training to effectively balance multimodal learning. The learning rate is dynamically adjusted during training to balance the speed of learning from each modality by the multimodal model, aiming for enhanced performance in both multimodal and unimodal predictions. We extensively evaluate MILES on four multimodal joint fusion tasks and compare its performance to seven state-of-the-art baselines. Our results show that MILES outperforms all baselines across all tasks and fusion methods considered in our study, effectively balancing modality usage during training. This results in improved multimodal performance and stronger modality encoders, which can be leveraged when dealing with unimodal samples or absent modalities. Overall, our work highlights the impact of balancing multimodal learning on improving model performance.