TSR · The desk

Science · September 2026

Articles in this category in September 2026.

• TSR Desk · science · 30 September 2026, 01:00 UTC

  Bison: Cross-Dataset Learning for Unseen-Compound Perturbation Prediction

• TSR Desk · science · 30 September 2026, 01:00 UTC

  Reliability Engineering for AI Systems: Challenges, Methods, and Directions

• TSR Desk · science · 30 September 2026, 01:00 UTC

  Cross Modality Image Translation In Medical Imaging Using Generative Frameworks

• TSR Desk · science · 30 September 2026, 01:00 UTC

  REALIS: A Curated Dataset for Studying the Challenges of AI Image Detection

• TSR Desk · science · 30 September 2026, 01:00 UTC

  Failure-Transparent Agents: Benchmarking Post-Failure Reporting in Tool-Using Language Models

• TSR Desk · science · 30 September 2026, 01:00 UTC

  Improving Large Language Models for Code through Runtime Program-State Reasoning

• TSR Desk · science · 30 September 2026, 01:00 UTC

  FOCUS: Benchmarking Retinal Model Generalization from Foundation Vision Encoders to Multimodal

• TSR Desk · science · 30 September 2026, 01:00 UTC

  GUI-GenBench: Evaluating Image Generation Models as Interactive GUI Environments

• TSR Desk · science · 30 September 2026, 01:00 UTC

  VehDyn: A Driving World Model Benchmark for Vehicle Dynamics

• TSR Desk · science · 30 September 2026, 01:00 UTC

  Laya as a Typed Probabilistic Assessor: An Independent Reproduction and a Preregistered Study

• TSR Desk · science · 30 September 2026, 01:00 UTC

  Timeline-Bench: Evaluating Agents on Realistic Video-Editing Tasks, from Raw Footage to Final

• TSR Desk · science · 30 September 2026, 01:00 UTC

  Measurement-Error-Aware Causal Distributed-Lag Quantile Modeling of Indoor Air Pollution and

• TSR Desk · science · 30 September 2026, 01:00 UTC

  The Undetected Damage of Quantization on Retrieval and How to Fix It

• TSR Desk · science · 30 September 2026, 01:00 UTC

  AquaWAM: A Dynamics-aware World Action Model for Underwater Embodied Agents

• TSR Desk · science · 30 September 2026, 01:00 UTC

  Rethinking Circuit Evaluation: Do Circuits Explain Model Errors?

• TSR Desk · science · 29 September 2026, 01:00 UTC

  Attribution Bias in Large Language Models

• TSR Desk · science · 29 September 2026, 01:00 UTC

  From Reward Signal to Visual Utility: A Controlled Audit of Medical VLM Post-Training

• TSR Desk · science · 29 September 2026, 01:00 UTC

  Backbone-Adaptive Evidence Routing for Robust Pairwise LLM Judging

• TSR Desk · science · 29 September 2026, 01:00 UTC

  Towards Unified Dynamic Face Landmark Detection

• TSR Desk · science · 29 September 2026, 01:00 UTC

  A Mechanistic Study of AI-Text Detection Neurons in Frozen BERT: Sparse Probing and Activation

• TSR Desk · science · 29 September 2026, 01:00 UTC

  Too Sure to Be Safe: Model Calibration for Reliable Log Anomaly Detection

• TSR Desk · science · 29 September 2026, 01:00 UTC

  Robust to Which Model Change? A Unified Evaluation of Robust Counterfactual Explanations

• TSR Desk · science · 29 September 2026, 01:00 UTC

  AcuityBench: Evaluating Clinical Acuity Identification and Uncertainty Alignment

• TSR Desk · science · 29 September 2026, 01:00 UTC

  UltraG-Bench: A Multi-task Benchmark for assessing Large Vision-Language Models on Pixel-level

• TSR Desk · science · 29 September 2026, 01:00 UTC

  GT-HarmBench: Benchmarking AI Safety Risks Through the Lens of Game Theory

• TSR Desk · science · 29 September 2026, 01:00 UTC

  PriceBench: A Diagnostic Benchmark for Price, Quality, and Brand Preferences in LLM Booking

• TSR Desk · science · 29 September 2026, 01:00 UTC

  BabelCoder: Agentic Code Translation with Specification Alignment

• TSR Desk · science · 29 September 2026, 01:00 UTC

  BioEVAL: A global, multi-institutional benchmark of large language and multimodal models for

• TSR Desk · science · 29 September 2026, 01:00 UTC

  SciR: A Controllable Benchmark for Scientific Reasoning in LLMs

• TSR Desk · science · 29 September 2026, 01:00 UTC

  Governed Deduction: Policy-Grounded Premise Authorization Beyond Relevance

• TSR Desk · science · 28 September 2026, 07:00 UTC

  DepthEvidence: Unifying Metric Depth Prediction and Geometric Reasoning in Multimodal Language

• TSR Desk · science · 28 September 2026, 07:00 UTC

  ScopeBench: Do Agents Preserve Engagement Boundaries Under Goal Pressure?

• TSR Desk · science · 28 September 2026, 07:00 UTC

  Models Got Talent: Identifying High Performing Wearable Human Activity Recognition Models

• TSR Desk · science · 28 September 2026, 07:00 UTC

  FlyAOC: Evaluating Agentic Ontology Curation of Drosophila Scientific Knowledge Bases

• TSR Desk · science · 28 September 2026, 07:00 UTC

  Geometric Inconsistency Localization in Multi-View Image Sets

• TSR Desk · science · 28 September 2026, 07:00 UTC

  Inquesto Score: A reliability Protocol For Voice Agents

• TSR Desk · science · 28 September 2026, 07:00 UTC

  SPO: Discovering Adaptive Large Neighborhood Search Operators via Stackelberg Program

• TSR Desk · science · 28 September 2026, 07:00 UTC

  SkillFlow: Scalable and Efficient Agent Skill Retrieval System

• TSR Desk · science · 28 September 2026, 07:00 UTC

  ReasonAudio: A Benchmark for Evaluating Reasoning Beyond Matching in Text-Audio Retrieval

• TSR Desk · science · 28 September 2026, 07:00 UTC

  LLM Parkinsonism: Executive-Control Failure, Token-Inefficient Persistence, and an

• TSR Desk · science · 28 September 2026, 07:00 UTC

  Stealth Apart, Harm Together: Skill Cascading Attacks on Skill-Based Agent Systems

• TSR Desk · science · 28 September 2026, 07:00 UTC

  SignTrace: Describe a Sign, Find the Word

• TSR Desk · science · 28 September 2026, 07:00 UTC

  Highlight-Then-Summarize: Learning to Compress Evidence for Long-Context Understanding

• TSR Desk · science · 28 September 2026, 07:00 UTC

  JevAdvBench: A Benchmark and Black-Box Attacks for Reinforcement Learning for Calibrated

• TSR Desk · science · 28 September 2026, 07:00 UTC

  ORCA: Evaluating LLMs on Data Science Code Translation

• TSR Desk · science · 26 September 2026, 01:00 UTC

  IatroBench: A Pre-Registered Benchmark of Clinical Omission in Language Models

• TSR Desk · science · 26 September 2026, 01:00 UTC

  ExplorationBench: Measuring AI Systems' Exploration in Verifiable Alien Worlds

• TSR Desk · science · 26 September 2026, 01:00 UTC

  Speculative Evaluation of Stochastic LLMs

• TSR Desk · science · 26 September 2026, 01:00 UTC

  Spot, Separate, and Enhance: Fully Generative Approach for Audio Mixing

• TSR Desk · science · 26 September 2026, 01:00 UTC

  RAPID: Robot Agentic Programming from Demonstrations

• TSR Desk · science · 26 September 2026, 01:00 UTC

  Synthetic Hospital: An Open, Verifiable, Physician-Validated Longitudinal EHR Benchmark

• TSR Desk · science · 26 September 2026, 01:00 UTC

  SheetMind: Actions Set Accuracy, Agents Set the Failure Mode

• TSR Desk · science · 26 September 2026, 01:00 UTC

  A Living Benchmark for Information Retrieval from Electronic Health Records

• TSR Desk · science · 26 September 2026, 01:00 UTC

  Wiring Beats Blending: Structure-Aware Compensation for Transformer Downscaling

• TSR Desk · science · 26 September 2026, 01:00 UTC

  To Trust or Not to Trust: Retrieval-Augmented Fact Checking in Speech

• TSR Desk · science · 26 September 2026, 01:00 UTC

  EnigmaForge: The Question Is Hidden in the Story

• TSR Desk · science · 26 September 2026, 01:00 UTC

  Baszta: Data-Centric Fine-Tuning of a Polish Multi-Label Safety Classifier

• TSR Desk · science · 25 September 2026, 01:00 UTC

  Deep Reinforcement Learning on Item-Compatibility Graphs for One-Dimensional Bin Packing

• TSR Desk · science · 25 September 2026, 01:00 UTC

  AgenticSizing: A Large Language Model-based Multi-Agent Framework for Analog Circuit Sizing

• TSR Desk · science · 25 September 2026, 01:00 UTC

  BixBench3: Benchmarking AI agents on research-study-scale computational biology tasks

• TSR Desk · science · 25 September 2026, 01:00 UTC

  Glucose-ML: A collection of longitudinal diabetes datasets for development of robust AI

• TSR Desk · science · 25 September 2026, 01:00 UTC

  Benchmarking Neural Defend ARCAS 1B: A Foundational Multimodal Deepfake Detection Model

• TSR Desk · science · 25 September 2026, 01:00 UTC

  The Sirens' Song: When Proximal Background Context Overshadows Distant Evidence

• TSR Desk · science · 25 September 2026, 01:00 UTC

  MGRL-RSCC: Multi-Granularity Reward Reinforcement Learning for Fine-Grained Remote Sensing

• TSR Desk · science · 25 September 2026, 01:00 UTC

  ArticleMiner: Ontology-Guided Knowledge Graph Construction from Scientific Publications

• TSR Desk · science · 25 September 2026, 01:00 UTC

  Trains but Doesn't Learn: A Post-Training Delivery Benchmark for LLM Agents as Forward-Deployed

• TSR Desk · science · 25 September 2026, 01:00 UTC

  ScholarStack: Layered Research Asset Orchestration and Cross-Task Reuse for Scientific Agents

• TSR Desk · science · 25 September 2026, 01:00 UTC

  A JEPA Recipe for Tabular Foundation Models

• TSR Desk · science · 25 September 2026, 01:00 UTC

  DolphinBench: Mapping the Pareto Frontier of Agent Memory

• TSR Desk · science · 25 September 2026, 01:00 UTC

  BigO(Bench): Can LLMs Generate Code with Controlled Time and Space Complexity?

• TSR Desk · science · 25 September 2026, 01:00 UTC

  Seeing the imagined: latent functional alignment in visual imagery decoding from fMRI data

• TSR Desk · science · 25 September 2026, 01:00 UTC

  Adversarial Course-of-Action Generation: Game-Theoretic Multi-Agent Algorithms for COA matching

• TSR Desk · science · 24 September 2026, 01:00 UTC

  WebArxiv: A Reproducible Benchmark for Evaluating Multimodal Web Agents on arXiv Tasks

• TSR Desk · science · 24 September 2026, 01:00 UTC

  RankCert: When Can Simulated Learners Safely Select an AI Tutor? Robust Decision Certification

• TSR Desk · science · 24 September 2026, 01:00 UTC

  eXplaining to Learn (eX2L): Regularization Using Contrastive Visual Explanation Pairs for

• TSR Desk · science · 24 September 2026, 01:00 UTC

  Refusal without Discrimination: What Encoded Prompts Do to Safety-Trained Models

• TSR Desk · science · 24 September 2026, 01:00 UTC

  Queer inclusion in speech datasets: An audit and taxonomy of practical tensions

• TSR Desk · science · 24 September 2026, 01:00 UTC

  Agent Memory: Characterization and System Implications of Stateful Long-Horizon Workloads

• TSR Desk · science · 24 September 2026, 01:00 UTC

  FinRED: An Expert-Guided Benchmark Generation and Evaluation Framework for Financial LLM

• TSR Desk · science · 24 September 2026, 01:00 UTC

  EADC: Evaluation of Advanced and Deep-level Compliance in Large Language Models

• TSR Desk · science · 24 September 2026, 01:00 UTC

  Testing-Driven Reliability Audit of Trajectory-Based Early Outcome Prediction for LLM Agents:

• TSR Desk · science · 24 September 2026, 01:00 UTC

  The Delegation Blind Spot: Auditing Product Decisions from Agent Choices

• TSR Desk · science · 24 September 2026, 01:00 UTC

  Terminal Shrinkage Averaging Reveals a Schedule-Estimator Interaction in LLM Pretraining

• TSR Desk · science · 24 September 2026, 01:00 UTC

  Grow the Harness, Not the Context: From Strategy-Free Scaffolds to Reusable Specialist Agents

• TSR Desk · science · 24 September 2026, 01:00 UTC

  Passes Alone, Fails Together: Benchmarking Semantic Coordination in Parallel LLM-Agent

• TSR Desk · science · 23 September 2026, 01:00 UTC

  Teaching LLMs to Self-Evolve: Cultivating Core Meta-Skills with Reinforcement Learning

• TSR Desk · science · 23 September 2026, 01:00 UTC

  Geometry of Values: Task Vector Composition for Ethical Preference Alignment in Language Models

• TSR Desk · science · 23 September 2026, 01:00 UTC

  What Should We Ask Next? Retrieval-Aware Question Learning under Partial Evidence

• TSR Desk · science · 23 September 2026, 01:00 UTC

  2nd Place Solution to the HANDS 2026 Workshop Challenge-Dexterous Grasp Motion Track:

• TSR Desk · science · 23 September 2026, 01:00 UTC

  Deep Learning-Enhanced Real-Time Wi-Fi Sensing Through Single Transceiver Pair

• TSR Desk · science · 23 September 2026, 01:00 UTC

  Trustworthy FinAInce: Unpacking How AI-Mediated Financial Advice is Judged

• TSR Desk · science · 23 September 2026, 01:00 UTC

  MOSCOPT: Mixture-of-Skills Collective Optimization for LLM Agents

• TSR Desk · science · 23 September 2026, 01:00 UTC

  LoRA Enhanced Contrastive Learning with SAS Vision Transformers

• TSR Desk · science · 23 September 2026, 01:00 UTC

  AutoRecLab: Describe the Experiment, Get the Code!

• TSR Desk · science · 23 September 2026, 01:00 UTC

  PACE: Precise AI Cinematic Expression

• TSR Desk · science · 23 September 2026, 01:00 UTC

  NeuSOGA3D: A Neuro-Symbolic Framework for Explainable 3D Geometric Reconstruction

• TSR Desk · science · 23 September 2026, 01:00 UTC

  Benchmarking the Explanatory Quality of Open-Weight Vision-Language Models in Face Recognition

• TSR Desk · science · 23 September 2026, 01:00 UTC

  Intent-Governed Tool Authorization for AI Agents

• TSR Desk · science · 23 September 2026, 01:00 UTC

  A Training-Free Proactive Defense Against Partial Speech Manipulation via Self-Embedding

• TSR Desk · science · 22 September 2026, 01:00 UTC

  Towards the Vision-Sound-Language-Action Paradigm: The HEAR Framework for Sound-Centric