STaR-DRO: Stateful Tsallis Reweighting for Group-Robust Structured Prediction

• URL: http://arxiv.org/abs/2604.09737v1
• Date: Thu, 09 Apr 2026 21:58:42 GMT
• Title: STaR-DRO: Stateful Tsallis Reweighting for Group-Robust Structured Prediction
• Abstract summary: We present a two-part framework for controllable inference and robust fine-tuning.<n>First, we introduce a task-agnostic prompting strategy to address format drift, label ambiguity, evidence hallucination, and metadata-conditioned confusion.<n>Second, we introduce STaR-DRO, a stateful robust optimization method for group heterogeneity.
• Score: 0.20165644958637055
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Structured prediction requires models to generate ontology-constrained labels, grounded evidence, and valid structure under ambiguity, label skew, and heterogeneous group difficulty. We present a two-part framework for controllable inference and robust fine-tuning. First, we introduce a task-agnostic prompting strategy that combines XML-based instruction structure, disambiguation rules, verification-style reasoning, schema constraints, and self-validation to address format drift, label ambiguity, evidence hallucination, and metadata-conditioned confusion in in-context structured generation. Second, we introduce STaR-DRO, a stateful robust optimization method for group heterogeneity. It combines Tsallis mirror descent with momentum-smoothed, centered group-loss signals and bounded excess-only multipliers so that only persistently hard groups above a neutral baseline are upweighted, concentrating learning where it is most needed while avoiding volatile, dense exponentiated-gradient reweighting and unnecessary loss from downweighting easier groups. We evaluate the combined framework on EPPC Miner, a benchmark for extracting hierarchical labels and evidence spans from patient-provider secure messages. Prompt engineering improves zero-shot by +15.44 average F1 across Code, Sub-code, and Span over four Llama models. Building on supervised fine-tuning, STaR-DRO further improves the hardest semantic decisions: on Llama-3.3-70B-Instruct, Code F1 rises from 79.24 to 81.47 and Sub-code F1 from 67.78 to 69.30, while preserving Span performance and reducing group-wise validation cross-entropy by up to 29.6% on the most difficult clinical categories. Because these rare and difficult groups correspond to clinically consequential communication behaviors, these gains are not merely statistical improvements: they directly strengthen communication mining reliability for patient-centered care analysis.

Related papers

• Imputation Meets Clustering: Exploiting Latent Subgroup Structure for Missing Data Recovery [52.29866804948133]
  We propose CAGI (Cluster-Aware Generative Imputation), a framework that reformulates clustering and imputation as a mutually reinforcing co-optimization process.<n>Experiments on 14 benchmark datasets with 15 representative baselines demonstrate the superiority of CAGI.
  arXiv  Detail & Related papers  (2026-07-08T02:50:02Z)
• Adaptive Group-Based Counterfactual Explanations for Time-Series Rehabilitation Data [17.023825093545582]
  We propose a two-stage framework for group-based counterfactual generation in high-dimensional IMU data.<n>We first show that Shapley-Adaptive perturbation (SA) group ranking preserves counterfactual validity but fails to enforce group-level sparsity.<n>We then introduce Learnable Gate (LG) methods, which incorporate trainable per-group relevance gates jointly optimized with masks.
  arXiv  Detail & Related papers  (2026-07-02T08:02:05Z)
• Dual-Branch Cross-Projection Debiasing through Diffusion-based Disentanglement [66.67463557274358]
  Foundation models trained on biased datasets often rely on spurious correlations between target labels and non-causal attributes.<n>We introduce Confidence-guided Concept Mining (CBCM) to identify reliable spurious attributes without attribute annotations.<n>We propose Dual-branch Cross-projection Debiasing (DCD), a prompt-tuning framework that separates target and spurious representations into two branches.
  arXiv  Detail & Related papers  (2026-06-23T05:28:47Z)
• Fuzzy-Geometric Branch-Point Modeling for Structure-Aware Augmentation of Handwritten Chinese Characters [2.7561351786180937]
  This paper proposes a fuzzy geometry-driven structure-aware (FGSA) augmentation framework.<n>We model branch points as fuzzy sets within the skeleton space, constructing a continuous branch-point membership field.<n>Experiments on CASIA-HWDB1.1, ChiSig, and LZUSig demonstrate that FGSA significantly reduces the word-level error rate.
  arXiv  Detail & Related papers  (2026-06-17T08:04:23Z)
• Too Correct to Learn: Reinforcement Learning on Saturated Reasoning Data [55.84428098924793]
  We propose Constrained Uniform Top-K Sampling (CUTS), a parameter-free decoding strategy enforcing structure-preserving exploration.<n>We integrate this into Mixed-CUTS, a training framework synergizing exploitative and exploratory rollouts to amplify intra-group advantage variance.<n> Notably, Mixed-CUTS improves Pass@1 accuracy on the challenging AIME25 benchmark by up to 15.1% over standard GRPO.
  arXiv  Detail & Related papers  (2026-04-20T16:43:28Z)
• Dynamic analysis enhances issue resolution [53.50448142467294]
  DAIRA (Dynamic Analysis-enhanced Issue Resolution Agent) is an automated repair framework that embeds dynamic analysis into the agent's reasoning cycle.<n> Driven by a Test Tracing-Driven methodology, DAIRA utilizes lightweight monitors to extract critical runtime data.<n>When powered by Gemini 3 Flash Preview, DAIRA establishes a new state-of-the-art (SOTA) performance, achieving a 79.4% resolution rate on the SWE-bench Verified dataset.
  arXiv  Detail & Related papers  (2026-03-23T14:48:54Z)
• Prompt Group-Aware Training for Robust Text-Guided Nuclei Segmentation [6.329891529118214]
  We reformulate prompt sensitivity as a group-wise consistency problem.<n>We show that our method improves Dice by an average of 2.16 points on six zero-shot cross-dataset tasks.
  arXiv  Detail & Related papers  (2026-03-06T15:34:08Z)
• Group Distributionally Robust Optimization-Driven Reinforcement Learning for LLM Reasoning [45.86058898829962]
  Multi-Ad Distributionally Robust Optimization (GDRO) is an optimization-first framework that moves beyond uniform reasoning.<n>We propose two independent GDRO games for post-training: Prompt-GDRO, which employs an EMA-debiased multiplicative-weight bandit sampler to target the intensive difficulty margin and upweight persistently hard groups without frequency bias; and Rollout-GDRO, which uses a shadow-price controller to reallocate rollouts across groups, maximizing gradient variance reduction on hard tasks under a fixed mean budget (compute-neutral)<n>We validate our framework on the DAPO 14.1k dataset using Q
  arXiv  Detail & Related papers  (2026-01-27T07:10:41Z)
• Chunking, Retrieval, and Re-ranking: An Empirical Evaluation of RAG Architectures for Policy Document Question Answering [0.0]
  The integration of Large Language Models (LLMs) into the public health policy sector offers a transformative approach to navigating the vast repositories of regulatory guidance maintained by agencies such as the Centers for Disease Control and Prevention (CDC)<n>The propensity for LLMs to generate hallucinations, defined as plausible but factually incorrect assertions, presents a critical barrier to the adoption of these technologies in high-stakes environments where information integrity is non-negotiable.<n>This empirical evaluation explores the effectiveness of Retrieval-Augmented Generation (RAG) architectures in mitigating these risks by grounding generative outputs in authoritative document context.
  arXiv  Detail & Related papers  (2026-01-21T20:52:48Z)
• GRCF: Two-Stage Groupwise Ranking and Calibration Framework for Multimodal Sentiment Analysis [20.77940776708036]
  Pairwise ordinal learning frameworks capture relative order by learning from comparisons.<n>They assign uniform importance to all comparisons, failing to adaptively focus on hard-to-rank samples.<n>We propose a Two-Stage Group-wise Ranking and Framework (GRCF) that adapts the philosophy of Group Relative Policy Optimization.<n>GRCF achieves state-of-the-art performance on core regression benchmarks, while also showing strong generalizability in classification tasks.
  arXiv  Detail & Related papers  (2026-01-14T16:26:44Z)
• Mitigating Clever Hans Strategies in Image Classifiers through Generating Counterexamples [15.618934546058277]
  Group distributional robustness methods rely on explicit group labels to upweight underrepresented groups.<n>We propose Counterfactual Knowledge Distillation (CFKD), a framework that generates diverse counterfactuals.<n>We demonstrate CFKD's efficacy across five datasets, spanning synthetic tasks to an industrial application.
  arXiv  Detail & Related papers  (2025-10-20T13:22:57Z)
• Retrieval is Not Enough: Enhancing RAG Reasoning through Test-Time Critique and Optimization [58.390885294401066]
  Retrieval-augmented generation (RAG) has become a widely adopted paradigm for enabling knowledge-grounded large language models (LLMs)<n>RAG pipelines often fail to ensure that model reasoning remains consistent with the evidence retrieved, leading to factual inconsistencies or unsupported conclusions.<n>We propose AlignRAG, a novel iterative framework grounded in Critique-Driven Alignment (CDA)<n>We introduce AlignRAG-auto, an autonomous variant that dynamically terminates refinement, removing the need to pre-specify the number of critique iterations.
  arXiv  Detail & Related papers  (2025-04-21T04:56:47Z)
• Modeling the Q-Diversity in a Min-max Play Game for Robust Optimization [61.39201891894024]
  Group distributionally robust optimization (group DRO) can minimize the worst-case loss over pre-defined groups. We reformulate the group DRO framework by proposing Q-Diversity. Characterized by an interactive training mode, Q-Diversity relaxes the group identification from annotation into direct parameterization.
  arXiv  Detail & Related papers  (2023-05-20T07:02:27Z)
• Examining and Combating Spurious Features under Distribution Shift [94.31956965507085]
  We define and analyze robust and spurious representations using the information-theoretic concept of minimal sufficient statistics. We prove that even when there is only bias of the input distribution, models can still pick up spurious features from their training data. Inspired by our analysis, we demonstrate that group DRO can fail when groups do not directly account for various spurious correlations.
  arXiv  Detail & Related papers  (2021-06-14T05:39:09Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.