Advantage Shaping as Surrogate Reward Maximization: Unifying Pass@K Policy Gradients

This note reconciles two seemingly distinct approaches to policy gradient optimization for the Pass@K objective in reinforcement learning with verifiable rewards: (1) direct REINFORCE-style methods, and (2) advantage-shaping techniques that directly modify GRPO. We show that these are two sides of the same coin. By reverse-engineering existing advantage-shaping algorithms, we reveal that they implicitly optimize surrogate rewards. We specifically interpret practical ``hard-example up-weighting'' modifications to GRPO as reward-level regularization. Conversely, starting from surrogate reward objectives, we provide a simple recipe for deriving both existing and new advantage-shaping methods. This perspective provides a lens for RLVR policy gradient optimization beyond our original motivation of Pass@K.

Efficient Forward-Only Data Valuation for Pretrained LLMs and VLMs

Quantifying the influence of individual training samples is essential for enhancing the transparency and accountability of large language models (LLMs) and vision-language models (VLMs). However, existing data valuation methods often rely on Hessian information or model retraining, making them computationally prohibitive for billion-parameter models. In this work, we introduce For-Value, a forward-only data valuation framework that enables scalable and efficient influence estimation for both LLMs and VLMs. By leveraging the rich representations of modern foundation models, For-Value computes influence scores using a simple closed-form expression based solely on a single forward pass, thereby eliminating the need for costly gradient computations. Our theoretical analysis demonstrates that For-Value accurately estimates per-sample influence by capturing alignment in hidden representations and prediction errors between training and validation samples. Extensive experiments show that For-Value matches or outperforms gradient-based baselines in identifying impactful fine-tuning examples and effectively detecting mislabeled data.

On the Effect of Negative Gradient in Group Relative Deep Reinforcement Optimization

Reinforcement learning (RL) has become popular in enhancing the reasoning capabilities of large language models (LLMs), with Group Relative Policy Optimization (GRPO) emerging as a widely used algorithm in recent systems. Despite GRPO's widespread adoption, we identify a previously unrecognized phenomenon we term Lazy Likelihood Displacement (LLD), wherein the likelihood of correct responses marginally increases or even decreases during training. This behavior mirrors a recently discovered misalignment issue in Direct Preference Optimization (DPO), attributed to the influence of negative gradients. We provide a theoretical analysis of GRPO's learning dynamic, identifying the source of LLD as the naive penalization of all tokens in incorrect responses with the same strength. To address this, we develop a method called NTHR, which downweights penalties on tokens contributing to the LLD. Unlike prior DPO-based approaches, NTHR takes advantage of GRPO's group-based structure, using correct responses as anchors to identify influential tokens. Experiments on math reasoning benchmarks demonstrate that NTHR effectively mitigates LLD, yielding consistent performance gains across models ranging from 0.5B to 3B parameters.

MedReason: Eliciting Factual Medical Reasoning Steps in LLMs via Knowledge Graphs

Medical tasks such as diagnosis and treatment planning require precise and complex reasoning, particularly in life-critical domains. Unlike mathematical reasoning, medical reasoning demands meticulous, verifiable thought processes to ensure reliability and accuracy. However, there is a notable lack of datasets that provide transparent, step-by-step reasoning to validate and enhance the medical reasoning ability of AI models. To bridge this gap, we introduce MedReason, a large-scale high-quality medical reasoning dataset designed to enable faithful and explainable medical problem-solving in large language models (LLMs). We utilize a structured medical knowledge graph (KG) to convert clinical QA pairs into logical chains of reasoning, or ``thinking paths'', which trace connections from question elements to answers via relevant KG entities. Each path is validated for consistency with clinical logic and evidence-based medicine. Our pipeline generates detailed reasoning for various medical questions from 7 medical datasets, resulting in a dataset of 32,682 question-answer pairs, each with detailed, step-by-step explanations. Experiments demonstrate that fine-tuning with our dataset consistently boosts medical problem-solving capabilities, achieving significant gains of up to 7.7% for DeepSeek-Ditill-8B. Our top-performing model, MedReason-8B, outperforms the Huatuo-o1-8B, a state-of-the-art medical reasoning model, by up to 4.2% on the clinical benchmark MedBullets. We also engage medical professionals from diverse specialties to assess our dataset's quality, ensuring MedReason offers accurate and coherent medical reasoning. Our data, models, and code will be publicly available.

Can Textual Gradient Work in Federated Learning?

Recent studies highlight the promise of LLM-based prompt optimization, especially with TextGrad, which automates differentiation'' via texts and backpropagates textual feedback. This approach facilitates training in various real-world applications that do not support numerical gradient propagation or loss calculation. In this paper, we systematically explore the potential and challenges of incorporating textual gradient into Federated Learning (FL). Our contributions are fourfold. Firstly, we introduce a novel FL paradigm, Federated Textual Gradient (FedTextGrad), that allows clients to upload locally optimized prompts derived from textual gradients, while the server aggregates the received prompts. Unlike traditional FL frameworks, which are designed for numerical aggregation, FedTextGrad is specifically tailored for handling textual data, expanding the applicability of FL to a broader range of problems that lack well-defined numerical loss functions. Secondly, building on this design, we conduct extensive experiments to explore the feasibility of FedTextGrad. Our findings highlight the importance of properly tuning key factors (e.g., local steps) in FL training. Thirdly, we highlight a major challenge in FedTextGrad aggregation: retaining essential information from distributed prompt updates. Last but not least, in response to this issue, we improve the vanilla variant of FedTextGrad by providing actionable guidance to the LLM when summarizing client prompts by leveraging the Uniform Information Density principle. Through this principled study, we enable the adoption of textual gradients in FL for optimizing LLMs, identify important issues, and pinpoint future directions, thereby opening up a new research area that warrants further investigation.

DARE the Extreme: Revisiting Delta-Parameter Pruning For Fine-Tuned Models

Storing open-source fine-tuned models separately introduces redundancy and increases response times in applications utilizing multiple models. Delta-parameter pruning (DPP), particularly the random drop and rescale (DARE) method proposed by Yu et al., addresses this by pruning the majority of delta parameters--the differences between fine-tuned and pre-trained model weights--while typically maintaining minimal performance loss. However, DARE fails when either the pruning rate or the magnitude of the delta parameters is large. We highlight two key reasons for this failure: (1) an excessively large rescaling factor as pruning rates increase, and (2) high mean and variance in the delta parameters. To push DARE's limits, we introduce DAREx (DARE the eXtreme), which features two algorithmic improvements: (1) DAREx-q, a rescaling factor modification that significantly boosts performance at high pruning rates (e.g., >30 % on COLA and SST2 for encoder models, with even greater gains in decoder models), and (2) DAREx-L2, which combines DARE with AdamR, an in-training method that applies appropriate delta regularization before DPP. We also demonstrate that DAREx-q can be seamlessly combined with vanilla parameter-efficient fine-tuning techniques like LoRA and can facilitate structural DPP. Additionally, we revisit the application of importance-based pruning techniques within DPP, demonstrating that they outperform random-based methods when delta parameters are large. Through this comprehensive study, we develop a pipeline for selecting the most appropriate DPP method under various practical scenarios.

Universal Debiased Editing on Foundation Models for Fair Medical Image Classification

In the era of Foundation Models' (FMs) rising prominence in AI, our study addresses the challenge of biases in medical images while using FM API, particularly spurious correlations between pixels and sensitive attributes. Traditional methods for bias mitigation face limitations due to the restricted access to web-hosted FMs and difficulties in addressing the underlying bias encoded within the FM API. We propose an U(niversal) D(ebiased) E(diting) strategy, termed UDE, which generates UDE noise to mask such spurious correlation. UDE is capable of mitigating bias both within the FM API embedding and the images themselves. Furthermore, UDE is suitable for both white-box and black-box FM APIs, where we introduced G(reedy) (Z)eroth-O(rder) (GeZO) optimization for it when the gradient is inaccessible in black-box APIs. Our whole pipeline enables fairness-aware image editing that can be applied across various medical contexts without requiring direct model manipulation or significant computational resources. Our empirical results demonstrate the method's effectiveness in maintaining fairness and utility across different patient groups and diseases. In the era of AI-driven medicine, this work contributes to making healthcare diagnostics more equitable, showcasing a practical solution for bias mitigation in pre-trained image FMs.

Content Conditional Debiasing for Fair Text Embedding

Mitigating biases in machine learning models has gained increasing attention in Natural Language Processing (NLP). Yet, only a few studies focus on fair text embeddings, which are crucial yet challenging for real-world applications. In this paper, we propose a novel method for learning fair text embeddings. We achieve fairness while maintaining utility trade-off by ensuring conditional independence between sensitive attributes and text embeddings conditioned on the content. Specifically, we enforce that embeddings of texts with different sensitive attributes but identical content maintain the same distance toward the embedding of their corresponding neutral text. Furthermore, we address the issue of lacking proper training data by using Large Language Models (LLMs) to augment texts into different sensitive groups. Our extensive evaluations demonstrate that our approach effectively improves fairness while preserving the utility of embeddings, representing a pioneering effort in achieving conditional independence for fair text embeddings.

Heterogeneous Federated Learning with Group-Aware Prompt Tuning

Transformers have achieved remarkable success in various machine-learning tasks, prompting their widespread adoption. In this paper, we explore their application in the context of federated learning (FL), with a particular focus on heterogeneous scenarios where individual clients possess diverse local datasets. To meet the computational and communication demands of FL, we leverage pre-trained Transformers and use an efficient prompt-tuning strategy. Our strategy introduces the concept of learning both shared and group prompts, enabling the acquisition of universal knowledge and group-specific knowledge simultaneously. Additionally, a prompt selection module assigns personalized group prompts to each input, aligning the global model with the data distribution of each client. This approach allows us to train a single global model that can automatically adapt to various local client data distributions without requiring local fine-tuning. In this way, our proposed method effectively bridges the gap between global and personalized local models in Federated Learning and surpasses alternative approaches that lack the capability to adapt to previously unseen clients. The effectiveness of our approach is rigorously validated through extensive experimentation and ablation studies.

Clinical-Inspired Cytological Whole Slide Image Screening with Just Slide-Level Labels

Cytology test is effective, non-invasive, convenient, and inexpensive for clinical cancer screening. ThinPrep, a commonly used liquid-based specimen, can be scanned to generate digital whole slide images (WSIs) for cytology testing. However, WSIs classification with gigapixel resolutions is highly resource-intensive, posing significant challenges for automated medical image analysis. In order to circumvent this computational impasse, existing methods emphasize learning features at the cell or patch level, typically requiring labor-intensive and detailed manual annotations, such as labels at the cell or patch level. Here we propose a novel automated Label-Efficient WSI Screening method, dubbed LESS, for cytology-based diagnosis with only slide-level labels. Firstly, in order to achieve label efficiency, we suggest employing variational positive-unlabeled (VPU) learning, enhancing patch-level feature learning using WSI-level labels. Subsequently, guided by the clinical approach of scrutinizing WSIs at varying fields of view and scales, we employ a cross-attention vision transformer (CrossViT) to fuse multi-scale patch-level data and execute WSI-level classification. We validate the proposed label-efficient method on a urine cytology WSI dataset encompassing 130 samples (13,000 patches) and FNAC 2019 dataset with 212 samples (21,200 patches). The experiment shows that the proposed LESS reaches 84.79%, 85.43%, 91.79% and 78.30% on a urine cytology WSI dataset, and 96.53%, 96.37%, 99.31%, 94.95% on FNAC 2019 dataset in terms of accuracy, AUC, sensitivity and specificity. It outperforms state-of-the-art methods and realizes automatic cytology-based bladder cancer screening.