12d ago
Learning General Representation of 12-Lead Electrocardiogram with a Joint-Embedding Predictive Architecture
Self-supervised learning approach for ECG signal representation using masked modeling from unlabeled medical data.
12d ago
Mitigating Extrinsic Gender Bias for Bangla Classification Tasks
Investigation of gender bias in Bangla language models with benchmark datasets for sentiment analysis, toxicity detection, hate speech, and sarcasm.
12d ago
OmniPrism: Learning Disentangled Visual Concept for Image Generation
Method for learning disentangled visual concepts in image generation to improve multi-aspect creative generation while reducing concept confusion.
12d ago
Neurons Speak in Ranges: Breaking Free from Discrete Neuronal Attribution
Analysis of polysemanticity in LLMs revealing neurons exhibit multiple semantic meanings, challenging discrete neuron attribution for model interpretation.
12d ago
AgentSociety: Large-Scale Simulation of LLM-Driven Generative Agents Advances Understanding of Human Behaviors and Society
Research on large-scale simulation of LLM-driven generative agents for studying human behavior and social dynamics through computational approaches.
12d ago
Constraining Sequential Model Editing with Editing Anchor Compression
Sequential model editing method with editing anchor compression to constrain parameter drift and maintain LLM general abilities during knowledge updates.
12d ago
Revitalizing Black-Box Interpretability: Actionable Interpretability for LLMs via Proxy Models
Budget-friendly proxy model framework for post-hoc interpretability of LLMs, enabling actionable explanations for prompt engineering and optimization.
12d ago
Chain-of-Zoom: Extreme Super-Resolution via Scale Autoregression and Preference Alignment
Autoregressive super-resolution framework decomposing extreme upsampling into intermediate scales with preference alignment for improved scalability.
12d ago
Gen-n-Val: Agentic Image Data Generation and Validation
Agentic framework for synthetic image data generation and validation addressing data scarcity and label noise in vision tasks like detection and segmentation.
12d ago
Enhancing the Safety of Medical Vision-Language Models by Synthetic Demonstrations
Safety enhancement for medical vision-language models using synthetic demonstrations to improve rejection of harmful clinical queries.
12d ago
Listener-Rewarded Thinking in VLMs for Image Preferences
Listener-rewarded thinking approach using reinforcement learning to train robust reward models for generative text-to-image and video models.
12d ago
Provable Post-Training Quantization: Theoretical Analysis of OPTQ and Qronos
Theoretical analysis providing quantitative guarantees for post-training quantization methods OPTQ and Qronos applied to LLMs and neural networks.
12d ago
VSI: Visual Subtitle Integration for Keyframe Selection to enhance Long Video Understanding
Keyframe selection method using visual subtitles for improved long video understanding with multimodal LLMs under context length constraints.
12d ago
Mitigating Domain Drift in Multi Species Segmentation with DINOv2: A Cross-Domain Evaluation in Herbicide Research Trials
DINOv2-based segmentation framework for plant species and damage detection in herbicide trials, addressing domain drift across real-world conditions.
12d ago
Investigating Multimodal Large Language Models to Support Usability Evaluation
Investigation of multimodal LLMs for automating usability evaluation of user interfaces by analyzing visual UI context and textual instructions.
12d ago
AR-KAN: Autoregressive-Weight-Enhanced Kolmogorov-Arnold Network for Time Series Forecasting
Kolmogorov-Arnold Network variant with autoregressive weights for time series forecasting, comparing performance against LLMs and ARIMA.
12d ago
STCast: Adaptive Boundary Alignment for Global and Regional Weather Forecasting
Spatial-temporal weather forecasting model with adaptive boundary alignment for improved global and regional predictions.
12d ago
On-the-Fly Adaptation to Quantization: Configuration-Aware LoRA for Efficient Fine-Tuning of Quantized LLMs
Configuration-aware LoRA adaptation for efficient fine-tuning of quantized LLMs on heterogeneous edge devices with privacy preservation.
12d ago
Adaptive Planning for Multi-Attribute Controllable Summarization with Monte Carlo Tree Search
Monte Carlo Tree Search approach for multi-attribute controllable summarization without per-attribute fine-tuning, enabling flexible constraint satisfaction.
12d ago
Traj2Action: A Co-Denoising Framework for Trajectory-Guided Human-to-Robot Skill Transfer
Co-denoising framework for transferring manipulation skills from human videos to robots by bridging morphological differences.
12d ago
Unmasking Puppeteers: Leveraging Biometric Leakage to Disarm Impersonation in AI-based Videoconferencing
Security research on defending AI-based videoconferencing systems against pose-expression latent hijacking attacks using biometric detection.
12d ago
Webscale-RL: Automated Data Pipeline for Scaling RL Data to Pretraining Levels
Automated pipeline for scaling reinforcement learning datasets to pretraining scale, addressing data bottleneck in RL for LLM training.
12d ago
Dejavu: Towards Experience Feedback Learning for Embodied Intelligence
Post-deployment learning framework for Vision-Language-Action policies using retrieved execution memories to improve embodied agent performance.
12d ago
RESample: A Robust Data Augmentation Framework via Exploratory Sampling for Robotic Manipulation
Data augmentation framework for robotic manipulation using Vision-Language-Action models to improve learning from limited demonstration datasets.
12d ago
LLM4Delay: Flight Delay Prediction via Cross-Modality Adaptation of Large Language Models and Aircraft Trajectory Representation
LLM-based framework for predicting flight delays using textual aeronautical information and aircraft trajectory data for air traffic management.
12d ago
How Similar Are Grokipedia and Wikipedia? A Multi-Dimensional Textual and Structural Comparison
Computational analysis comparing 17,790 articles between Grokipedia (AI-generated) and Wikipedia examining textual and structural biases.
12d ago
EGMOF: Efficient Generation of Metal-Organic Frameworks Using a Hybrid Diffusion-Transformer Architecture
EGMOF: hybrid diffusion-transformer for metal-organic framework generation with inverse design capabilities for materials discovery.
12d ago
Evolutionary Optimization Trumps Adam Optimization on Embedding Space Exploration
Inference-time optimization using evolutionary algorithms on prompt embeddings for diffusion model control without fine-tuning.
12d ago
Structured Uncertainty guided Clarification for LLM Agents
Structured uncertainty framework for LLM agents with tool-calling to generate principled clarifying questions for ambiguous user instructions.
12d ago
Commanding Humanoid by Free-form Language: A Large Language Action Model with Unified Motion Vocabulary
Language-conditioned humanoid robot control using LLM with unified motion vocabulary for free-form command execution and embodied AI.
12d ago
Bharat Scene Text: A Novel Comprehensive Dataset and Benchmark for Indian Language Scene Text Understanding
Bharat Scene Text dataset and benchmark for Indian language scene text recognition addressing script diversity and font variations.
12d ago
See, Hear, and Understand: Benchmarking Audiovisual Human Speech Understanding in Multimodal Large Language Models
AV-SpeakerBench: multimodal LLM benchmark with 3,212 questions evaluating audiovisual speech understanding and speaker-speech alignment in video.
12d ago
From Navigation to Refinement: Revealing the Two-Stage Nature of Flow-based Diffusion Models through Oracle Velocity
Analysis of flow-based diffusion models revealing two-stage behavior through oracle velocity field computation and memorization-generalization tradeoffs.
12d ago
Out-of-the-box: Black-box Causal Attacks on Object Detectors
Research on adversarial perturbations for object detectors using black-box attacks to expose vulnerabilities and understand attack mechanisms.
12d ago
SkillFactory: Self-Distillation For Learning Cognitive Behaviors
Research on self-distillation methods for teaching language models to leverage cognitive skills like verification and backtracking without base model exposure.
12d ago
Relational Visual Similarity
Research on relational visual similarity in computer vision showing how humans perceive analogical relationships beyond attribute similarity.
12d ago
Multi-agent Adaptive Mechanism Design
Framework combining mechanism design and online learning for sequential mechanism design where principal learns agent beliefs while ensuring truthfulness.
12d ago
The Two-Stage Decision-Sampling Hypothesis: Understanding the Emergence of Self-Reflection in RL-Trained LLMs
Mechanistic study of self-reflection emergence in RL-trained LLMs, proposing two-stage decision-sampling hypothesis to explain unified optimization producing distinct capabilities.
12d ago
Adversarial Evasion Attacks on Computer Vision using SHAP Values
White-box adversarial attack method on computer vision models using SHAP values to generate imperceptible evasion attacks.
12d ago
Screen, Cache, and Match: A Training-Free Causality-Consistent Reference Frame Framework for Human Animation
Training-free framework for human video animation using cached reference frames to model long-range dependencies while preserving temporal coherence.
12d ago
On the Limits of Layer Pruning for Generative Reasoning in Large Language Models
Analysis showing layer pruning of LLMs degrades generative reasoning tasks beyond surface degradation, causing loss of algorithmic capabilities.
12d ago
Tiled Prompts: Overcoming Prompt Misguidance in Image and Video Super-Resolution
Method addressing prompt misguidance in diffusion-based super-resolution by using tiled prompts for localized semantic guidance.
12d ago
SPEAR: An Engineering Case Study of Multi-Agent Coordination for Smart Contract Auditing
Multi-agent framework for smart contract auditing using specialized agents for planning, execution, and recovery with coordination protocols.
12d ago
Overstating Attitudes, Ignoring Networks: LLM Biases in Simulating Misinformation Susceptibility
Study demonstrating LLM biases when simulating misinformation susceptibility, showing models overstate attitudes and ignore network effects present in humans.
12d ago
Exploring Teachers' Perspectives on Using Conversational AI Agents for Group Collaboration
Qualitative study of 33 K12 teachers' perspectives on using conversational AI agents to scaffold group collaboration in classrooms.
12d ago
An Adaptive Model Selection Framework for Demand Forecasting under Horizon-Induced Degradation to Support Business Strategy and Operations
Adaptive framework for demand forecasting model selection addressing horizon-induced performance degradation in inventory planning.
12d ago
SubQuad: Near-Quadratic-Free Structure Inference with Distribution-Balanced Objectives in Adaptive Receptor framework
Pipeline combining subquadratic retrieval and GPU-accelerated kernels for analyzing immune repertoires at population scale.
12d ago
Descriptor: Parasitoid Wasps and Associated Hymenoptera Dataset (DAPWH)
Dataset of parasitoid wasps and hymenoptera for taxonomic identification and biodiversity monitoring.
12d ago
Reinforcement-aware Knowledge Distillation for LLM Reasoning
Knowledge distillation method for distilling RL-trained LLMs with chain-of-thought reasoning into smaller student models while preserving reasoning capabilities.
12d ago
Why Adam Can Beat SGD: Second-Moment Normalization Yields Sharper Tails
Theoretical analysis explaining why Adam optimizer outperforms SGD through second-moment normalization using stopping-time and martingale analysis.