Papers

New on arXiv

Thu 13 Feb 2025

373 papers

Digi-Q: Learning Q-Value Functions for Training Device-Control AgentsLLM agents and tool use → GUI and mobile agents
Knowledge Integration Strategies in Autonomous Vehicle Prediction and Planning: A Comprehensive SurveyEnd-to-end autonomous driving → Reasoning VLA and VLM driving policies
Diverse Transformer Decoding for Offline Reinforcement Learning Using Financial Algorithmic ApproachesOffline RL and imitation learning from data → Decision transformers and sequence modeling
On the robustness of multimodal language model towards distractionsMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
Statistical Coherence Alignment for Large Language Model Representation Learning Through Tensor Field ConvergenceLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
AgentGuard: Repurposing Agentic Orchestrator for Safety Evaluation of Tool OrchestrationLLM agents and tool use → Agent threats, guardrails and governance
Noise Controlled CT Super-Resolution with Conditional Diffusion ModelDiffusion models → Medical imaging and segmentation
TableTalk: Scaffolding Spreadsheet Development with a Language AgentLLM agents and tool use → Financial trading and forecasting agents
CellFlux: Simulating Cellular Morphology Changes via Flow MatchingDiffusion models → Flow matching, time series and science
Non-Markovian Discrete Diffusion with Causal Language ModelsDiffusion models → Discrete and text diffusion
Jailbreak Attack Initializations as Extractors of Compliance DirectionsLLM safety and security → Jailbreaking
QueryAttack: Jailbreaking Aligned Large Language Models Using Structured Non-natural Query LanguageLLM alignment / RLHF / preference optimisation → Safety alignment and jailbreak attacks +1
ZeroBench: An Impossible Visual Benchmark for Contemporary Large Multimodal ModelsMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
The Hidden Dimensions of LLM Alignment: A Multi-Dimensional Analysis of Orthogonal Safety DirectionsLLM reasoning → Reasoning for agents, planning and code +2
Are Smarter LLMs Safer? Exploring Safety-Reasoning Trade-offs in Prompting and Fine-TuningLLM reasoning → Reasoning for agents, planning and code +1
Theoretical Benefit and Limitation of Diffusion Language ModelDiffusion models → Discrete and text diffusion
MME-CoT: Benchmarking Chain-of-Thought in Large Multimodal Models for Reasoning Quality, Robustness, and EfficiencyLLM reasoning → Multimodal reasoning +1
Exploring the Potential of Encoder-free Architectures in 3D LMMsMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
Variational Rectified Flow MatchingDiffusion models → Flow matching, time series and science
Score-of-Mixture Training: Training One-Step Generative Models Made Simple via Score Estimation of Mixture DistributionsDiffusion models → Diffusion foundations and sampling
SelfCite: Self-Supervised Alignment for Context Attribution in Large Language ModelsLLM alignment / RLHF / preference optimisation → Preference alignment for recommenders and RAG
CoT-Valve: Length-Compressible Chain-of-Thought TuningLLM reasoning → Test-time compute and efficient thinking
GAIA: A Global, Multi-modal, Multi-scale Vision-Language Dataset for Remote Sensing Image AnalysisMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
KIMAs: A Configurable Knowledge Integrated Multi-Agent SystemRetrieval-augmented generation → Agentic RAG, agents and memory +1
Diffusing DeBias: Synthetic Bias Amplification for Model DebiasingDiffusion models → Diffusion foundations and sampling
EmbodiedBench: Comprehensive Benchmarking Multi-modal Large Language Models for Vision-Driven Embodied AgentsMultimodal LLMs / vision-language models → Embodied and 3D vision-language models +1
Long-Term TalkingFace Generation via Motion-Prior Conditional Diffusion ModelDiffusion models → Audio, speech and music
Diffusion Models for Molecules: A Survey of Methods and TasksDiffusion models → Flow matching, time series and science
When and How Does CLIP Enable Domain and Compositional Generalization?Multimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP
Eidetic Learning: an Efficient and Provable Solution to Catastrophic ForgettingMixture-of-experts → Mixture of LoRA experts
A Survey of Multilingual Reasoning in Language ModelsLLM reasoning → Reasoning benchmarks and evaluation
Pixel-Level Reasoning Segmentation via Multi-turn ConversationsLLM reasoning → Multimodal reasoning
Redistribute Ensemble Training for Mitigating Memorization in Diffusion ModelsDiffusion models → Diffusion foundations and sampling
On multi-token prediction for efficient LLM inferenceEfficient LLM inference and serving → Speculative decoding
ImageRAG: Dynamic Image Retrieval for Reference-Guided Image GenerationDiffusion models → Image generation, text-to-image and editing
A hierarchical approach for assessing the vulnerability of tree-based classification models to membership inference attackLLM safety and security → Membership inference attacks
SQuARE: Sequential Question Answering Reasoning Engine for Enhanced Chain-of-Thought in Large Language ModelsLLM reasoning → Prompting and chain-of-thought reasoning
Language Agents as Digital Representatives in Collective Decision-MakingLLM agents and tool use → Social simulation and multi-agent debate
ThunderServe: High-performance and Cost-efficient LLM Serving in Cloud EnvironmentsEfficient LLM inference and serving → Serving systems and request scheduling
Non-asymptotic Analysis of Diffusion Annealed Langevin Monte Carlo for Generative ModellingDiffusion models → Diffusion foundations and sampling
KET-RAG: A Cost-Efficient Multi-Granular Indexing Framework for Graph-RAGRetrieval-augmented generation → Knowledge-graph and structured-knowledge RAG
When Do Neural Networks Learn World Models?World models → World models for language agents
ConsistentDreamer: View-Consistent Meshes Through Balanced Multi-View Gaussian OptimizationDiffusion models → 3D, point cloud and scene generation
GEVRM: Goal-Expressive Video Generation Model For Robust Visual ManipulationVision-language-action models → Generalist VLA models and benchmarks
From large language models to multimodal AI: A scoping review on the potential of generative AI in medicineDiffusion models → Medical imaging and segmentation
Reliable Conversational Agents under ASP Control that Understand Natural LanguageLLM agents and tool use → General LLM agent frameworks and benchmarks
Logical Lease Litigation: Prolog and LLMs for Rental Law Compliance in New YorkLLM reasoning → Reasoning for agents, planning and code
Thinking beyond the anthropomorphic paradigm benefits LLM researchLLM reasoning → Reasoning benchmarks and evaluation
RefineCoder: Iterative Improving of Large Language Models via Adaptive Critique Refinement for Code GenerationLLM reasoning → Reasoning for agents, planning and code
FLAME: Flexible LLM-Assisted Moderation EngineLLM safety and security → Jailbreaking
LimSim Series: An Autonomous Driving Simulation Platform for Validation and EnhancementEnd-to-end autonomous driving → Generative simulation for driving
E-MD3C: Taming Masked Diffusion Transformers for Efficient Zero-Shot Object CustomizationDiffusion models → Diffusion foundations and sampling
Regularization can make diffusion models more efficientDiffusion models → Diffusion foundations and sampling
Can Vision-Language Models Infer Speaker's Ignorance? The Role of Visual and Linguistic CuesMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
A Survey of LLM Alignment: Instruction Understanding, Intention Reasoning, and Reliable GenerationLLM reasoning → Reasoning benchmarks and evaluation +1
Logical Reasoning in Large Language Models: A SurveyLLM reasoning → Reasoning benchmarks and evaluation
Enhancing RAG with Active Learning on Conversation Records: Reject Incapables and Answer CapablesRetrieval-augmented generation → Conversational and speech RAG
StyleBlend: Enhancing Style-Specific Content Creation in Text-to-Image Diffusion ModelsDiffusion models → Image generation, text-to-image and editing
CRANE: Reasoning with constrained LLM generationLLM reasoning → Reasoning benchmarks and evaluation
Vision-Language In-Context Learning Driven Few-Shot Visual Inspection ModelMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
Adapting Language-Specific LLMs to a Reasoning Model in One Day via Model Merging -- An Open RecipeLLM reasoning → Reasoning benchmarks and evaluation
AIDE: Agentically Improve Visual Language Model with Domain ExpertsMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
Typhoon T1: An Open Thai Reasoning ModelLLM reasoning → Reasoning benchmarks and evaluation
MTDP: A Modulated Transformer based Diffusion Policy ModelDiffusion models → Diffusion and flow policies for robotics
Mechanistic Unveiling of Transformer Circuits: Self-Influence as a Key to Model ReasoningLLM reasoning → Prompting and chain-of-thought reasoning
Beyond Shallow Behavior: Task-Efficient Value-Based Multi-Task Offline MARL via Skill DiscoveryOffline RL and imitation learning from data → Offline RL algorithms
Text-driven 3D Human Generation via Contrastive Preference OptimizationLLM alignment / RLHF / preference optimisation → Diffusion and multimodal model alignment
Topo2Seq: Enhanced Topology Reasoning via Topology Sequence LearningLLM reasoning → Chain-of-thought reasoning for driving
RTBAS: Defending LLM Agents Against Prompt Injection and Privacy LeakageLLM agents and tool use → Agent threats, guardrails and governance +1
Medicine on the Edge: Comparative Performance Analysis of On-Device LLMs for Clinical ReasoningLLM reasoning → Clinical and biomedical applications
On the Promise for Assurance of Differentiable Neurosymbolic Reasoning ParadigmsLLM reasoning → Reasoning benchmarks and evaluation
Dynamic watermarks in images generated by diffusion modelsDiffusion models → Robustness, privacy and adversarial aspects +1
CopySpec: Accelerating LLMs with Speculative Copy-and-Paste Without Compromising QualityEfficient LLM inference and serving → Speculative decoding
Self-Consistency of the Internal Reward Models Improves Self-Rewarding Language ModelsLLM reasoning → Reinforcement learning for reasoning +2
PathFinder: A Multi-Modal Multi-Agent System for Medical Diagnostic Decision-Making Applied to HistopathologyLLM agents and tool use → Medical and scientific domain agents
Diffusion Models Through a Global Lens: Are They Culturally Inclusive?Diffusion models → Image generation, text-to-image and editing
InfiniteHiP: Extending Language Model Context Up to 3 Million Tokens on a Single GPUEfficient LLM inference and serving → KV cache optimization
3D-Grounded Vision-Language Framework for Robotic Task Planning: Automated Prompt Synthesis and Supervised ReasoningLLM reasoning → Spatial and embodied reasoning +1
EnigmaEval: A Benchmark of Long Multimodal Reasoning ChallengesLLM reasoning → Reasoning benchmarks and evaluation
Application Modernization with LLMs: Addressing Core Challenges in Reliability, Security, and Quality
Toward Total Recall: Enhancing FAIRness through AI-Driven Metadata Standardization
Cause-effect perception in an object place task
DreamLLM-3D: Affective Dream Reliving using Large Language Model and 3D Generative AI
A 28 nm AI microcontroller with tightly coupled zero-standby power weight memory featuring standard logic compatible 4 Mb 4-bits/cell embedded flash technology
Russo-Ukrainian war disinformation detection in suspicious Telegram channels
Brain-Inspired Exploration of Functional Networks and Key Neurons in Large Language Models
MC2SleepNet: Multi-modal Cross-masking with Contrastive Learning for Sleep Stage Classification
PixleepFlow: A Pixel-Based Lifelog Framework for Predicting Sleep Quality and Stress Level
AIvaluateXR: An Evaluation Framework for on-Device AI in XR with Benchmarking Results
TRKM: Twin Restricted Kernel Machines for Classification and Regression
A Survey of QUD Models for Discourse Processing
nanoML for Human Activity Recognition
Application-oriented automatic hyperparameter optimization for spiking neural network prototyping
GoRA: Gradient-driven Adaptive Low Rank Adaptation
MUDDFormer: Breaking Residual Bottlenecks in Transformers via Multiway Dynamic Dense Connections
Antimatter Annihilation Vertex Reconstruction with Deep Learning for ALPHA-g Radial Time Projection Chamber
CFIRSTNET: Comprehensive Features for Static IR Drop Estimation with Neural Network
TastepepAI, An artificial intelligence platform for taste peptide de novo design
Chronic Diseases Prediction Using ML
SinSim: Sinkhorn-Regularized SimCLR

All 373