Papers

New on arXiv

Thu 10 Oct 2024

419 papers

The Geometry of Concepts: Sparse Autoencoder Feature StructureMechanistic interpretability → Sparse autoencoder training and evaluation
Towards Next-Generation LLM-based Recommender Systems: A Survey and BeyondLLM reasoning → LLM reasoning for recommendation
TPO: Aligning Large Language Models with Multi-branch & Multi-step Preference TreesLLM reasoning → Mathematical reasoning +1
Diversity of Thought Elicits Stronger Reasoning Capabilities in Multi-Agent Debate FrameworksLLM reasoning → Reasoning for agents, planning and code +1
Prompt Engineering a Schizophrenia Chatbot: Utilizing a Multi-Agent Approach for Enhanced Compliance with Prompt InstructionsLLM agents and tool use → Medical and scientific domain agents
A Framework for Collaborating a Large Language Model Tool in Brainstorming for Triggering Creative ThoughtsLLM reasoning → Reasoning benchmarks and evaluation
REDO: Execution-Free Runtime Error Detection for COding AgentsLLM agents and tool use → Software engineering coding agents
Catastrophic Cyber Capabilities Benchmark (3CB): Robustly Evaluating LLM Agent Cyber Offense CapabilitiesLLM agents and tool use → Agent threats, guardrails and governance
IceDiff: High Resolution and High-Quality Sea Ice Forecasting with Generative Diffusion PriorDiffusion models → Flow matching, time series and science
Bilinear MLPs enable weight-based mechanistic interpretabilityMechanistic interpretability → Circuits and mechanistic analysis
What is Left After Distillation? How Knowledge Transfer Impacts Fairness and BiasKnowledge distillation → General knowledge distillation methods
AgroGPT: Efficient Agricultural Vision-Language Model with Expert TuningMultimodal LLMs / vision-language models → Agricultural vision-language models
KV Prediction for Improved Time to First TokenEfficient LLM inference and serving → KV cache optimization
KnowGraph: Knowledge-Enabled Anomaly Detection via Logical Reasoning on Graph DataLLM reasoning → Knowledge- and tool-augmented reasoning
Optimizing Vital Sign Monitoring in Resource-Constrained Maternal Care: An RL-Based Restless Bandit ApproachDeep reinforcement learning methods → RL for clinical treatment and surgery
Exploring Natural Language-Based Strategies for Efficient Number Learning in Children through Reinforcement LearningDeep reinforcement learning methods → RL training of language and agent models
Agents Thinking Fast and Slow: A Talker-Reasoner ArchitectureLLM reasoning → Reasoning for agents, planning and code +1
Evaluating Differentially Private Synthetic Data Generation in High-Stakes DomainsSynthetic data and self-improvement for LLMs → Privacy-preserving tabular data synthesis
Do You Know What You Are Talking About? Characterizing Query-Knowledge Relevance For Reliable Retrieval Augmented GenerationRetrieval-augmented generation → RAG pipeline design and evaluation
COS-DPO: Conditioned One-Shot Multi-Objective Fine-Tuning FrameworkLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
Avoiding mode collapse in diffusion models fine-tuned with reinforcement learningDiffusion models → Diffusion and flow policies for robotics +1
Swing-by Dynamics in Concept Learning and Compositional GeneralizationDiffusion models → Diffusion foundations and sampling
UNIQ: Offline Inverse Q-learning for Avoiding Undesirable DemonstrationsOffline RL and imitation learning from data → Imitation learning methods +1
Global Lyapunov functions: a long-standing open problem in mathematics, with symbolic transformersLLM reasoning → Mathematical reasoning
Increasing the Difficulty of Automatically Generated Questions via Reinforcement Learning with Synthetic PreferenceRetrieval-augmented generation → RAG robustness and adaptive retrieval
Meissonic: Revitalizing Masked Generative Transformers for Efficient High-Resolution Text-to-Image SynthesisDiffusion models → Image generation, text-to-image and editing
Flex-MoE: Modeling Arbitrary Modality Combination via the Flexible Mixture-of-ExpertsMixture-of-experts → Multimodal MoE models
LatteCLIP: Unsupervised CLIP Fine-Tuning via LMM-Synthetic TextsMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP
Emergent Visual Grounding in Large Multimodal Models Without Grounding SupervisionMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
DICE: Discrete Inversion Enabling Controllable Editing for Multinomial Diffusion and Masked Generative ModelsDiffusion models → Image generation, text-to-image and editing
Mono-InternVL: Pushing the Boundaries of Monolithic Multimodal Large Language Models with Endogenous Visual Pre-trainingMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
Efficient Dictionary Learning with Switch Sparse AutoencodersMechanistic interpretability → Sparse autoencoder training and evaluation
From Exploration to Mastery: Enabling LLMs to Master Tools via Self-Driven InteractionsLLM agents and tool use → Tool learning and function calling
MathCoder2: Better Math Reasoning from Continued Pretraining on Model-translated Mathematical CodeLLM reasoning → Mathematical reasoning
GenARM: Reward Guided Generation with Autoregressive Reward Model for Test-time AlignmentLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
HybridBooth: Hybrid Prompt Inversion for Efficient Subject-Driven GenerationDiffusion models → Image generation, text-to-image and editing
DifFRelight: Diffusion-Based Facial Performance RelightingDiffusion models → 3D, point cloud and scene generation
Scaling Laws For Diffusion TransformersDiffusion models → Diffusion foundations and sampling
MRAG-Bench: Vision-Centric Evaluation for Retrieval-Augmented Multimodal ModelsRetrieval-augmented generation → Multimodal RAG
Sample then Identify: A General Framework for Risk Control and Assessment in Multimodal Large Language ModelsMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
Chain-of-Sketch: Enabling Global Visual ReasoningLLM reasoning → Multimodal reasoning
Agent S: An Open Agentic Framework that Uses Computers Like a HumanLLM agents and tool use → GUI and mobile agents
DART: Denoising Autoregressive Transformer for Scalable Text-to-Image GenerationDiffusion models → Image generation, text-to-image and editing
Progressive Autoregressive Video Diffusion ModelsDiffusion models → Video generation
Rewarding Progress: Scaling Automated Process Verifiers for LLM ReasoningLLM reasoning → Reinforcement learning for reasoning +1
Insight Over Sight: Exploring the Vision-Knowledge Conflicts in Multimodal LLMsMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
DelTA: An Online Document-Level Translation Agent Based on Multi-Level MemoryLLM agents and tool use → General LLM agent frameworks and benchmarks
Steering Masked Discrete Diffusion Models via Discrete Denoising Posterior PredictionDiffusion models → Flow matching, time series and science
Think Beyond Size: Adaptive Prompting for More Effective ReasoningLLM reasoning → Prompting and chain-of-thought reasoning
Mars: Situated Inductive Reasoning in an Open-World EnvironmentLLM reasoning → Reasoning for agents, planning and code
Q-VLM: Post-training Quantization for Large Vision-Language ModelsMultimodal LLMs / vision-language models → Visual token pruning and efficiency
Optima: Optimizing Effectiveness and Efficiency for LLM-Based Multi-Agent SystemLLM agents and tool use → General LLM agent frameworks and benchmarks
A Closer Look at Machine Unlearning for Large Language ModelsLLM safety and security → Machine unlearning
What Makes Large Language Models Reason in (Multi-Turn) Code Generation?LLM reasoning → Reasoning for agents, planning and code
Can Knowledge Graphs Make Large Language Models More Trustworthy? An Empirical Study Over Open-ended Question AnsweringLLM reasoning → Knowledge- and tool-augmented reasoning
Unstable Unlearning: The Hidden Risk of Concept Resurgence in Diffusion ModelsDiffusion models → Robustness, privacy and adversarial aspects +1
Unlearning-based Neural InterpretationsLLM safety and security → Machine unlearning
Teaching-Inspired Integrated Prompting Framework: A Novel Approach for Enhancing Reasoning in Large Language ModelsLLM reasoning → Mathematical reasoning
Reward-Augmented Data Enhances Direct Preference Alignment of LLMsLLM alignment / RLHF / preference optimisation → LLM preference optimization methods +1
VerifierQ: Enhancing LLM Test Time Compute with Q-Learning-based VerifiersLLM reasoning → Reinforcement learning for reasoning
Divide and Translate: Compositional First-Order Logic Translation and Verification for Complex Logical ReasoningLLM reasoning → Reasoning benchmarks and evaluation
The Computational Complexity of Circuit Discovery for Inner InterpretabilityMechanistic interpretability → Circuits and mechanistic analysis
Probabilistic Satisfaction of Temporal Logic Constraints in Reinforcement Learning via Adaptive Policy-SwitchingDeep reinforcement learning methods → Interpretable and logic-guided RL
More Experts Than Galaxies: Conditionally-overlapping Experts With Biologically-Inspired Fixed RoutingMixture-of-experts → MoE theory and gating functions
Towards Synergistic, Generalized, and Efficient Dual-System for Robotic ManipulationVision-language-action models → Generalist VLA models and benchmarks
Neuroplastic Expansion in Deep Reinforcement LearningDeep reinforcement learning methods → Deep RL algorithms and representations
LADIMO: Face Morph Generation through Biometric Template Inversion with Latent DiffusionDiffusion models → Robustness, privacy and adversarial aspects
Omni-MATH: A Universal Olympiad Level Mathematic Benchmark For Large Language ModelsLLM reasoning → Mathematical reasoning
Offline Hierarchical Reinforcement Learning via Inverse OptimizationOffline RL and imitation learning from data → Offline RL algorithms
Boosting Hierarchical Reinforcement Learning with Meta-Learning for Complex Task AdaptationDeep reinforcement learning methods → Deep RL algorithms and representations
Generated Bias: Auditing Internal Bias Dynamics of Text-To-Image Generative ModelsDiffusion models → Image generation, text-to-image and editing
Constrained Skill Discovery: Quadruped Locomotion with Unsupervised Reinforcement LearningDeep reinforcement learning methods → RL for robot locomotion and manipulation
FDDM: Frequency-Decomposed Diffusion Model for Rectum Cancer Dose Prediction in RadiotherapyDiffusion models → Medical imaging and segmentation
Benchmarking Agentic Workflow GenerationLLM agents and tool use → Software engineering coding agents
RDT-1B: a Diffusion Foundation Model for Bimanual ManipulationDiffusion models → Diffusion and flow policies for robotics
HeGraphAdapter: Tuning Multi-Modal Vision-Language Models with Heterogeneous Graph AdapterMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
Autonomous Vehicles Path Planning under Temporal Logic SpecificationsEnd-to-end autonomous driving → Imitation-learning end-to-end planners
Semantic Self-Consistency: Enhancing Language Model Reasoning via Semantic WeightingLLM reasoning → Test-time compute and efficient thinking
Minority-Focused Text-to-Image Generation via Prompt OptimizationDiffusion models → Image generation, text-to-image and editing
Masked Generative Priors Improve World Models Sequence Modelling CapabilitiesWorld models → Latent dynamics models and model-based RL +1
Simple ReFlow: Improved Techniques for Fast Flow ModelsDiffusion models → Flow matching, time series and science
MGMD-GAN: Generalization Improvement of Generative Adversarial Networks with Multiple Generator Multiple Discriminator Framework Against Membership Inference AttacksLLM safety and security → Membership inference attacks
Mastering Contact-rich Tasks by Combining Soft and Rigid Robotics with Imitation LearningOffline RL and imitation learning from data → Robot manipulation imitation learning
Modeling User Preferences with Automatic Metrics: Creating a High-Quality Preference Dataset for Machine TranslationLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
On the grid-sampling limit SDEDeep reinforcement learning methods → RL theory and sample complexity
Dialectical Behavior Therapy Approach to LLM PromptingLLM reasoning → Prompting and chain-of-thought reasoning
HARIVO: Harnessing Text-to-Image Models for Video GenerationDiffusion models → Video generation
textitJump Your Steps: Optimizing Sampling Schedule of Discrete Diffusion ModelsDiffusion models → Diffusion foundations and sampling
Data Augmentation for Surgical Scene Segmentation with Anatomy-Aware Diffusion ModelsDiffusion models → Medical imaging and segmentation
StepTool: Enhancing Multi-Step Tool Usage in LLMs via Step-Grained Reinforcement LearningLLM agents and tool use → Reinforcement learning for LLM agents
SLIM: Let LLM Learn More and Forget Less with Soft LoRA and Identity MixtureMixture-of-experts → Mixture of LoRA experts
Hallo2: Long-Duration and High-Resolution Audio-Driven Portrait Image AnimationDiffusion models → Audio, speech and music
AgentBank: Towards Generalized LLM Agents via Fine-Tuning on 50000+ Interaction TrajectoriesLLM agents and tool use → General LLM agent frameworks and benchmarks
Autonomous Driving in Unstructured Environments: How Far Have We Come?End-to-end autonomous driving → Generative simulation for driving
The Power of Input: Benchmarking Zero-Shot Sim-To-Real Transfer of Reinforcement Learning Control Policies for Quadrotor ControlDeep reinforcement learning methods → RL for robot navigation and UAV control
Relational Diffusion Distillation for Efficient Image GenerationDiffusion models → Diffusion foundations and sampling
MACPO: Weak-to-Strong Alignment via Multi-Agent Contrastive Preference OptimizationLLM agents and tool use → Reinforcement learning for LLM agents +1
MotionAura: Generating High-Quality and Motion Consistent Videos using Discrete DiffusionDiffusion models → Video generation
Mechanistic Permutability: Match Features Across LayersMechanistic interpretability → Sparse autoencoder training and evaluation
FLIER: Few-shot Language Image Models Embedded with Latent RepresentationsMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP +1

All 419