3/19/2026
A Contextual Help Browser Extension to Assist Digital Illiterate Internet Users
Browser extension combining curated dictionary with OpenAI LLM to provide contextual help for technical terms via tooltips.
3/19/2026
rSDNet: Unified Robust Neural Learning against Label Noise and Adversarial Attacks
Neural network training robust to both label noise and adversarial attacks via unified loss function approach.
3/19/2026
Benchmarking Reinforcement Learning via Stochastic Converse Optimality: Generating Systems with Known Optimal Policies
Benchmarking framework for reinforcement learning algorithms using stochastic converse optimality with known optimal policies.
3/19/2026
Automated Grammar-based Algebraic Multigrid Design With Evolutionary Algorithms
Evolutionary algorithms to automatically design efficient multigrid cycles for solving partial differential equations.
3/19/2026
Interpretable Cross-Domain Few-Shot Learning with Rectified Target-Domain Local Alignment
Cross-domain few-shot learning with vision-language models like CLIP for fine-grained visual recognition tasks such as medical diagnosis.
3/19/2026
FINER: MLLMs Hallucinate under Fine-grained Negative Queries
Benchmark and analysis of hallucinations in multimodal LLMs with fine-grained negative queries covering multi-object, multi-attribute, and multi-relation scenarios.
3/19/2026
Post-Training Local LLM Agents for Linux Privilege Escalation with Verifiable Rewards
Post-training framework for small local LLM agents performing Linux privilege escalation tasks with verifiable rewards and resource constraints.
3/19/2026
Adaptive Guidance for Retrieval-Augmented Masked Diffusion Models
Adaptive guidance method for retrieval-augmented masked diffusion models that resolves conflicts between retrieved context and parametric knowledge.
3/19/2026
WeatherReasonSeg: A Benchmark for Weather-Aware Reasoning Segmentation in Visual Language Models
Benchmark for evaluating vision-language model reasoning and segmentation performance under adverse weather conditions with degraded visual cues.
3/19/2026
Can Blindfolded LLMs Still Trade? An Anonymization-First Framework for Portfolio Optimization
Framework for testing LLM trading agents with anonymized market data to validate genuine market understanding versus memorized ticker associations.
3/19/2026
Eye image segmentation using visual and concept prompts with Segment Anything Model 3 (SAM3)
Evaluation of Segment Anything Model 3 for eye image segmentation comparing text and visual prompting modes against prior SAM versions.
3/19/2026
Machine Learning for Network Attacks Classification and Statistical Evaluation of Machine Learning for Network Attacks Classification and Adversarial Learning Methodologies for Synthetic Data Generation
Survey of machine learning methods for network intrusion detection and adversarial learning techniques for synthetic attack data generation.
3/19/2026
SARE: Sample-wise Adaptive Reasoning for Training-free Fine-grained Visual Recognition
Training-free fine-grained visual recognition method using sample-wise adaptive reasoning with large vision-language models for subordinate-level category classification.
3/19/2026
Attention Sinks Induce Gradient Sinks
Analysis showing attention sinks in transformers induce gradient concentration during backpropagation under causal masking, affecting training dynamics.
3/19/2026
CoVerRL: Breaking the Consensus Trap in Label-Free Reasoning via Generator-Verifier Co-Evolution
LLM training approach using generator-verifier co-evolution to escape consensus trap and improve reasoning without ground-truth labels via reinforcement learning.
3/19/2026
A Dual Certificate Approach to Sparsity in Infinite-Width Shallow Neural Networks
Theoretical analysis of sparsity in infinite-width shallow ReLU networks trained with total variation regularization using duality theory.
3/19/2026
RangeAD: Fast On-Model Anomaly Detection
On-model anomaly detection method that leverages primary model representations to detect distributional shifts without separate AD models.
3/19/2026
EVA: Aligning Video World Models with Executable Robot Actions via Inverse Dynamics Rewards
Video world model framework with inverse dynamics rewards that ensures generated robot action sequences satisfy rigid-body and kinematic constraints.
3/19/2026
Fine-Grained Post-Training Quantization for Large Vision Language Models with Quantization-Aware Integrated Gradients
Post-training quantization method for large vision-language models using quantization-aware integrated gradients to reduce memory and computational overhead.
3/19/2026
Dropout Robustness and Cognitive Profiling of Transformer Models via Stochastic Inference
Study of dropout-induced variability and uncertainty in transformer models across 19 architectures using Monte Carlo sampling for inference-time evaluation.
3/19/2026
FailureMem: A Failure-Aware Multimodal Framework for Autonomous Software Repair
Multimodal framework for automated program repair using LLMs that jointly reasons over code, issue descriptions, and GUI screenshots to improve debugging workflows.
3/19/2026
CodeScout: An Effective Recipe for Reinforcement Learning of Code Search Agents
Reinforcement learning method for training code search agents to localize relevant files, classes, and functions in large repositories as prerequisite for coding tasks.
3/19/2026
Generative Control as Optimization: Time Unconditional Flow Matching for Adaptive and Robust Robotic Control
arXiv paper on inferring stage-play spatial layouts from narrative text, demonstrating language model spatial reasoning for automating dramaturgy.
3/19/2026
How do LLMs Compute Verbal Confidence
arXiv paper proposing GeCO, time-unconditional flow matching framework for adaptive robotic control using diffusion models.
3/19/2026
Procedural Generation of Algorithm Discovery Tasks in Machine Learning
arXiv paper investigating how LLMs compute verbal confidence scores and whether they're generated just-in-time or cached during inference.
3/19/2026
Mitigating LLM Hallucinations through Domain-Grounded Tiered Retrieval
arXiv paper introducing DiscoGen, procedural generator for algorithm discovery tasks addressing evaluation and contamination issues in ML benchmarks.
3/19/2026
Differential Attention-Augmented BiomedCLIP with Asymmetric Focal Optimization for Imbalanced Multi-Label Video Capsule Endoscopy Classification
arXiv paper proposing domain-grounded tiered retrieval architecture to reduce LLM hallucinations through systematic factual verification.
3/19/2026
RAMP: Reinforcement Adaptive Mixed Precision Quantization for Efficient On Device LLM Inference
Reinforcement learning framework for adaptive mixed-precision quantization of LLMs on resource-constrained devices with per-layer bit width optimization.
3/19/2026
scicode-lint: Detecting Methodology Bugs in Scientific Python Code with LLM-Generated Patterns
LLM-based linter detecting methodology bugs in scientific Python code that produce plausible but incorrect results.
3/19/2026
Differential Privacy in Generative AI Agents: Analysis and Optimal Tradeoffs
Analysis of privacy risks from enterprise data exposure in LLM-integrated systems with optimal differential privacy tradeoffs for agents.
3/19/2026
IndicSafe: A Benchmark for Evaluating Multilingual LLM Safety in South Asia
First systematic safety evaluation of LLMs across 12 Indic languages using 6,000 culturally grounded prompts in low-resource settings.
3/19/2026
CARE: Covariance-Aware and Rank-Enhanced Decomposition for Enabling Multi-Head Latent Attention
Technique for converting grouped-query attention to multi-head latent attention with improved expressivity and reduced KV-cache costs.
3/19/2026
VideoAtlas: Navigating Long-Form Video in Logarithmic Compute
Method for efficient long-form video processing with hierarchical grid representation enabling lossless, scalable video understanding.
3/19/2026
TDAD: Test-Driven Agentic Development - Reducing Code Regressions in AI Coding Agents via Graph-Based Impact Analysis
Open-source tool and benchmark for reducing code regressions in AI coding agents using AST-based impact analysis and test-driven development.
3/19/2026
Toward Scalable Automated Repository-Level Datasets for Software Vulnerability Detection
Automated approach for generating repository-level vulnerability detection datasets at scale beyond function-centric benchmarks.
3/19/2026
Loc3R-VLM: Language-based Localization and 3D Reasoning with Vision-Language Models
Framework augmenting 2D vision-language models with 3D spatial understanding and viewpoint-aware reasoning capabilities.
3/19/2026
Unified Spatio-Temporal Token Scoring for Efficient Video VLMs
Token pruning approach for efficient video vision-language models addressing temporal redundancy in video-based downstream tasks.
3/19/2026
Signal in the Noise: Polysemantic Interference Transfers and Predicts Cross-Model Influence
Study of polysemanticity in language models using sparse autoencoders to identify semantic interference patterns transferable across models.
3/19/2026
MLlm-DR: Towards Explainable Depression Recognition with MultiModal Large Language Models
Framework for explainable depression diagnosis using multimodal large language models on interview videos with depression score predictions.
3/19/2026
InPhyRe Discovers: Large Multimodal Models Struggle in Inductive Physical Reasoning
Evaluation showing large multimodal models struggle with inductive physical reasoning tasks beyond laws observed during training.
3/19/2026
See, Think, Act: Teaching Multimodal Agents to Effectively Interact with GUI by Identifying Toggles
Benchmark and analysis of multimodal agents' GUI interaction capabilities, identifying toggle control as a key bottleneck in graphical user interface automation.
3/19/2026
ScheduleMe: Multi-Agent Calendar Assistant
Multi-agent calendar assistant using graph-structured coordination with supervisory agent overseeing specialized task agents for natural language Google Calendar management.
3/19/2026
Assessing LLM Reasoning Through Implicit Causal Chain Discovery in Climate Discourse
Diagnostic evaluation of LLM reasoning capabilities through implicit causal chain discovery task, testing nine LLMs on mechanistic causal reasoning in climate discourse.
3/19/2026
TheraMind: A Strategic and Adaptive Agent for Longitudinal Psychological Counseling
Multi-agent LLM system for longitudinal psychological counseling with emotional understanding, adaptive strategies, and long-term memory capabilities.
3/19/2026
Efficient LLM Safety Evaluation through Multi-Agent Debate
Research on improving LLM safety evaluation using multi-agent debate with HAJailBench, a 11,100-sample human-annotated jailbreak benchmark across diverse attack methods.
3/19/2026
Safety-Preserving PTQ via Contrastive Alignment Loss
Safety-preserving post-training quantization via contrastive alignment loss to maintain behavioral safety during LLM model compression.
3/19/2026
Aligning Probabilistic Beliefs under Informative Missingness: LLM Steerability in Clinical Reasoning
Study on steering LLM probabilistic beliefs under informative missingness patterns for improved clinical reasoning with incomplete data.
3/19/2026
Stepwise Think-Critique: A Unified Framework for Robust and Interpretable LLM Reasoning
Stepwise Think-Critique framework unifying reasoning and verification in LLMs through intertwined critical thinking for robust problem-solving.
3/19/2026
CircuitLM: A Multi-Agent LLM-Aided Design Framework for Generating Circuit Schematics from Natural Language Prompts
CircuitLM multi-agent LLM pipeline generating circuit schematics from natural language, addressing hallucination and constraint violation issues in EDA.
3/19/2026
PaperScout: An Autonomous Agent for Academic Paper Search with Process-Aware Sequence-Level Policy Optimization
PaperScout autonomous agent for academic paper search using reinforcement learning to dynamically decide tool invocation for complex conditional queries.