Introduces the UCSF-PDGM-VQA dataset of 2387 QA pairs from 473 glioma MRI studies and demonstrates that state-of-the-art VLMs exhibit modality collapse on multi-sequence 3D medical images.
Title resolution pending
20 Pith papers cite this work. Polarity classification is still indexing.
20
Pith papers citing it
representative citing papers
Introduces the GeoDial dataset of 1.3K multimodal geometry tutoring dialogs grounded in diagram highlights, proposes an annotation protocol, and shows that fine-tuned VLMs improve dialog but struggle with accurate highlights.
Surprisal minimization over goal-directed alternatives generated by language models provides the strongest account of production choices in open-ended dialogue compared to uniform information density or length-based costs.
SG-RAG frames retrieval as subgraph matching to ensure LLMs meet every condition in factual queries and reports large gains over baselines on a new 120k-pair ERQA dataset.
VLMs as judges exhibit informativeness bias by favoring detailed but image-inconsistent answers; BIRCH mitigates it by first correcting answers against the image, reducing bias up to 17% and improving performance up to 9.8%.
SWE-RL uses RL on software evolution data to train LLMs achieving 41% on SWE-bench Verified with generalization to other reasoning tasks.
UniCoder applies symbolic attribute alignment via an auxiliary LLM and reference-guided optimization in RL to achieve SOTA visual-to-code generation on ChartMimic, UniSVG, Design2Code, and ScreenBench.
Interventions in LLM-simulated user experiments induce distribution shifts in latent attributes that create confounding bias, diagnosable with negative control outcomes and partially mitigated by adding setting-relevant persona details.
The paper proposes Strategic Prior-data Fitted Network (SPN), an inference-time framework that adapts pretrained tabular foundation models (PFNs) to strategic manipulation by aligning predictions with approximated post-manipulation distributions via strategic in-context examples.
AMARIS augments rubric updates in RL for LLMs with a persistent memory of rollout analyses and prior edits, yielding gains such as +2.8 points on GPQA-Diamond over local-adaptive baselines.
DMN achieves over 90% attack success rate on GPT-4o, Gemini-2.5-pro and Claude Sonnet 4 by distributing instructions, supplying multimodal evidence, and adding number chain tasks across multiple images.
PA-BDM adapts block diffusion by switching to causal intra-block denoising and dynamically committing reliable prefixes to KV cache, yielding higher accuracy and 71.6% higher throughput than a comparable baseline on document benchmarks.
Current audio-language models fail to use clinical multimodal context for dysarthric speech recognition, but context-aware LoRA fine-tuning delivers large accuracy gains on the SAP dataset.
DR-MMSearchAgent derives batch-wide trajectory advantages and uses differentiated Gaussian rewards to prevent premature collapse in multimodal agents, outperforming MMSearch-R1 by 8.4% on FVQA-test.
PDDL-Mind improves LLM accuracy on theory-of-mind benchmarks by over 5% by translating stories into verifiable PDDL states that decouple environment tracking from belief inference.
Filtering job posting data before LLM-assisted clustering and hierarchical labeling yields taxonomies with better AI skill coverage than unfiltered approaches.
TF-IDF identifies labeled experts in the top 25 recommendations 79.5% of the time versus 51.5% for GPT-4o mini on an astronomy observatory dataset.
R2IF uses a composite reward (binary correctness, CoT effectiveness, and parameter-level SMV) under GRPO to align LLM reasoning with function-calling decisions, improving accuracy and reasoning quality on BFCL/ACEBench.
SimpleSearch-VL improves Qwen3-VL multimodal agent baselines by 15.8-16 points on average using 7K total training examples and reaches parity with Gemini-3-Pro on the 30B variant.
Proposes Modality-Aware Credit Assignment (MoCA) with blindfolded-reasoning proxy to reward perception fidelity separately from reasoning in VLMs.
citing papers explorer
-
UCSF-PDGM-VQA: Visual Question Answering dataset for brain tumor MRI interpretation
Introduces the UCSF-PDGM-VQA dataset of 2387 QA pairs from 473 glioma MRI studies and demonstrates that state-of-the-art VLMs exhibit modality collapse on multi-sequence 3D medical images.
-
GeoDial: A Multimodal Conversational Tutoring Dataset for Geometry Problem-Solving with Visual Tutor Turns
Introduces the GeoDial dataset of 1.3K multimodal geometry tutoring dialogs grounded in diagram highlights, proposes an annotation protocol, and shows that fine-tuned VLMs improve dialog but struggle with accurate highlights.
-
Surprisal Minimisation over Goal-directed Alternatives Predicts Production Choice in Dialogue
Surprisal minimization over goal-directed alternatives generated by language models provides the strongest account of production choices in open-ended dialogue compared to uniform information density or length-based costs.
-
Structure Guided Retrieval-Augmented Generation for Factual Queries
SG-RAG frames retrieval as subgraph matching to ensure LLMs meet every condition in factual queries and reports large gains over baselines on a new 120k-pair ERQA dataset.
-
When Vision-Language Models Judge Without Seeing: Exposing Informativeness Bias
VLMs as judges exhibit informativeness bias by favoring detailed but image-inconsistent answers; BIRCH mitigates it by first correcting answers against the image, reducing bias up to 17% and improving performance up to 9.8%.
-
SWE-RL: Advancing LLM Reasoning via Reinforcement Learning on Open Software Evolution
SWE-RL uses RL on software evolution data to train LLMs achieving 41% on SWE-bench Verified with generalization to other reasoning tasks.
-
UniCoder: Unified Visual-to-Code Generation via Symbolic Rewards and Reference-Guided Code Optimization
UniCoder applies symbolic attribute alignment via an auxiliary LLM and reference-guided optimization in RL to achieve SOTA visual-to-code generation on ChartMimic, UniSVG, Design2Code, and ScreenBench.
-
The Illusion of Intervention: Your LLM-Simulated Experiment is an Observational Study
Interventions in LLM-simulated user experiments induce distribution shifts in latent attributes that create confounding bias, diagnosable with negative control outcomes and partially mitigated by adding setting-relevant persona details.
-
When Tabular Foundation Models Meet Strategic Tabular Data: A Prior Alignment Approach
The paper proposes Strategic Prior-data Fitted Network (SPN), an inference-time framework that adapts pretrained tabular foundation models (PFNs) to strategic manipulation by aligning predictions with approximated post-manipulation distributions via strategic in-context examples.
-
AMARIS: A Memory-Augmented Rubric Improvement System for Rubric-Based Reinforcement Learning
AMARIS augments rubric updates in RL for LLMs with a persistent memory of rollout analyses and prior edits, yielding gains such as +2.8 points on GPQA-Diamond over local-adaptive baselines.
-
DMN: A Compositional Framework for Jailbreaking Multimodal LLMs with Multi-Image Inputs
DMN achieves over 90% attack success rate on GPT-4o, Gemini-2.5-pro and Claude Sonnet 4 by distributing instructions, supplying multimodal evidence, and adding number chain tasks across multiple images.
-
Prefix-Adaptive Block Diffusion for Efficient Document Recognition
PA-BDM adapts block diffusion by switching to causal intra-block denoising and dynamically committing reliable prefixes to KV cache, yielding higher accuracy and 71.6% higher throughput than a comparable baseline on document benchmarks.
-
When Audio-Language Models Fail to Leverage Multimodal Context for Dysarthric Speech Recognition
Current audio-language models fail to use clinical multimodal context for dysarthric speech recognition, but context-aware LoRA fine-tuning delivers large accuracy gains on the SAP dataset.
-
DR-MMSearchAgent: Deepening Reasoning in Multimodal Search Agents
DR-MMSearchAgent derives batch-wide trajectory advantages and uses differentiated Gaussian rewards to prevent premature collapse in multimodal agents, outperforming MMSearch-R1 by 8.4% on FVQA-test.
-
PDDL-Mind: Large Language Models are Capable on Belief Reasoning with Reliable State Tracking
PDDL-Mind improves LLM accuracy on theory-of-mind benchmarks by over 5% by translating stories into verifiable PDDL states that decouple environment tracking from belief inference.
-
Building a Custom Taxonomy of AI Skills and Tasks from the Ground Up with Job Postings
Filtering job posting data before LLM-assisted clustering and hierarchical labeling yields taxonomies with better AI skill coverage than unfiltered approaches.
-
Traditional statistical representations outperform generative AI in identifying expert peer reviewers
TF-IDF identifies labeled experts in the top 25 recommendations 79.5% of the time versus 51.5% for GPT-4o mini on an astronomy observatory dataset.
-
R2IF: Aligning Reasoning with Decisions via Composite Rewards for Interpretable LLM Function Calling
R2IF uses a composite reward (binary correctness, CoT effectiveness, and parameter-level SMV) under GRPO to align LLM reasoning with function-calling decisions, improving accuracy and reasoning quality on BFCL/ACEBench.
-
SimpleSearch-VL: A Simple Recipe for Multimodal Agentic Deep Search
SimpleSearch-VL improves Qwen3-VL multimodal agent baselines by 15.8-16 points on average using 7K total training examples and reaches parity with Gemini-3-Pro on the 30B variant.
-
Bad Seeing or Bad Thinking? Rewarding Perception for Multimodal Reasoning
Proposes Modality-Aware Credit Assignment (MoCA) with blindfolded-reasoning proxy to reward perception fidelity separately from reasoning in VLMs.