RECAP: Transparent Inference-Time Emotion Alignment for Medical Dialogue Systems

• URL: http://arxiv.org/abs/2509.10746v1
• Date: Fri, 12 Sep 2025 23:30:45 GMT
• Title: RECAP: Transparent Inference-Time Emotion Alignment for Medical Dialogue Systems
• Abstract summary: Large language models in healthcare often miss critical emotional cues, delivering medically sound but emotionally flat advice.<n>We present RECAP, an inference-time framework that adds structured emotional reasoning without retraining.
• Score: 5.725572628136799
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Large language models in healthcare often miss critical emotional cues, delivering medically sound but emotionally flat advice. This is especially problematic in clinical contexts where patients are distressed and vulnerable, and require empathic communication to support safety, adherence, and trust. We present RECAP (Reflect-Extract-Calibrate-Align-Produce), an inference-time framework that adds structured emotional reasoning without retraining. By decomposing empathy into transparent appraisal-theoretic stages and exposing per-dimension Likert signals, RECAP produces nuanced, auditable responses. Across EmoBench, SECEU, and EQ-Bench, RECAP improves emotional reasoning by 22-28% on 8B models and 10-13% on larger models over zero-shot baselines. Clinician evaluations further confirm superior empathetic communication. RECAP shows that modular, theory-grounded prompting can systematically enhance emotional intelligence in medical AI while preserving the accountability required for deployment.

Related papers

• ER-EDF: A Psychology-Grounded Emotion Regulation Framework for Speech Empathetic Dialogue Generation in Large Audio-Language Models [13.58098743681316]
  Empathetic response generation in spoken dialogue systems requires both accurate emotion perception and appropriate emotion regulation.<n>We propose ER-EDF, a framework that explicitly decouples emotion perception and emotion regulation in large audio-language models.<n>Experiments show ER-EDF consistently improves empathetic response quality in both automatic and human evaluations.
  arXiv  Detail & Related papers  (2026-09-14T06:07:44Z)
• ENPMR-Bench: Benchmarking Proactive Memory Retrieval for Emotional Support Agents [53.83618283684421]
  We introduce ENPMR-Bench, a benchmark for evaluating Emotional Need-aware Proactive Memory Retrieval.<n> ENPMR-Bench includes over 1,800 memory-augmented dialogues and defines structured mappings between emotional needs and supportive memory types.
  arXiv  Detail & Related papers  (2026-05-26T16:22:35Z)
• EGI: A Multimodal Emotional AI Framework for Enhancing Scrum Master Real-time Self-Awareness [0.0]
  This paper proposes a novel application integrating four AI models to monitor the unconsciously expressed emotions of Scrum Masters and meeting organizers.<n>The system achieved an ASR word error rate WER of 10% in simulated meeting environments.
  arXiv  Detail & Related papers  (2026-05-17T22:55:20Z)
• EmoLLM: Appraisal-Grounded Cognitive-Emotional Co-Reasoning in Large Language Models [15.744777556347708]
  EmoLLM is an appraisal-grounded framework for IQ/EQ co-reasoning in dialogue.<n>We train EmoLLM in a multi-turn role-play environment with reinforcement learning.<n>EmoLLM improves emotional state outcomes and response quality over strong baselines.
  arXiv  Detail & Related papers  (2026-03-17T14:17:50Z)
• ADEPT: RL-Aligned Agentic Decoding of Emotion via Evidence Probing Tools -- From Consensus Learning to Ambiguity-Driven Emotion Reasoning [67.22219034602514]
  We introduce ADEPT (Agentic Decoding of Emotion via Evidence Probing Tools), a framework that reframes emotion recognition as a multi-turn inquiry process.<n> ADEPT transforms an SLLM into an agent that maintains an evolving candidate emotion set and adaptively invokes dedicated semantic and acoustic probing tools.<n>We show that ADEPT improves primary emotion accuracy in most settings while substantially improving minor emotion characterization.
  arXiv  Detail & Related papers  (2026-02-13T08:33:37Z)
• Technosocial risks of ideal emotion recognition technologies: A defense of the (social) value of emotional expressions [51.56484100374058]
  I argue that the appeal of such systems rests on a misunderstanding of the social functions emotional expression.<n>ERTs threaten this expressive space by collapsing epistemic friction, displacing meaning with technology-mediated affective profiles, and narrowing the space for aspirational and role-sensitive expressions.<n>I argue that, although it is intuitive to think that increasing accuracy would legitimise such systems, in the case of ERTs accuracy does not straightforwardly justify their deployment, and may, in some contexts, provide a reason for regulatory restraint.
  arXiv  Detail & Related papers  (2026-02-09T14:20:42Z)
• Reflecting Twice before Speaking with Empathy: Self-Reflective Alternating Inference for Empathy-Aware End-to-End Spoken Dialogue [53.95386201009769]
  We introduce EmpathyEval, a descriptive natural-language-based evaluation model for assessing empathetic quality in spoken dialogues.<n>We propose ReEmpathy, an end-to-end Spoken Language Models that enhances empathetic dialogue through a novel Empathetic Self-Reflective Alternating Inference mechanism.
  arXiv  Detail & Related papers  (2026-01-26T09:04:50Z)
• Empathy Applicability Modeling for General Health Queries [16.390464387095175]
  We introduce the Empathy Applicability Framework (EAF), a theory-driven approach that classifies patient queries in terms of the applicability of emotional reactions and interpretations.<n>EAF provides a framework for identifying empathy needs before response generation, establishes a benchmark for anticipatory empathy modeling.
  arXiv  Detail & Related papers  (2026-01-14T18:47:02Z)
• E-THER: A Multimodal Dataset for Empathic AI - Towards Emotional Mismatch Awareness [3.8298581733964903]
  E-THER is the first Person-Centered Therapy-grounded multimodal dataset with multidimensional annotations for verbal-visual incongruence detection.<n>We show that our incongruence-trained models outperform general-purpose models in critical traits.
  arXiv  Detail & Related papers  (2025-09-02T08:58:32Z)
• COMPEER: Controllable Empathetic Reinforcement Reasoning for Emotional Support Conversation [47.0476311232988]
  We propose controllable empathetic reasoning, which combines natural language reasoning with structured psychological steps.<n>We employ reinforcement learning with a unified process-outcome reward model that delivers precise feedback.<n>Our approach significantly improves model's emotional support ability, advancing the development of empathetic, human-like support systems.
  arXiv  Detail & Related papers  (2025-08-13T06:09:32Z)
• RLVER: Reinforcement Learning with Verifiable Emotion Rewards for Empathetic Agents [67.46032287312339]
  Large language models (LLMs) excel at logical and algorithmic reasoning, yet their emotional intelligence (EQ) still lags far behind their cognitive prowess.<n>We introduce RLVER, the first end-to-end reinforcement learning framework that leverages verifiable emotion rewards from simulated users.<n>Our results show that RLVER is a practical route toward emotionally intelligent and broadly capable language agents.
  arXiv  Detail & Related papers  (2025-07-03T18:33:18Z)
• Balancing Knowledge Delivery and Emotional Comfort in Healthcare Conversational Systems [24.901611078628527]
  We utilize a large language model to rewrite a real-world interactive medical dialogue dataset.<n>We generate patient queries with negative emotions and corresponding medical responses aimed at soothing the patient's emotions while addressing their concerns.
  arXiv  Detail & Related papers  (2025-06-16T16:54:03Z)
• IntentionESC: An Intention-Centered Framework for Enhancing Emotional Support in Dialogue Systems [74.0855067343594]
  In emotional support conversations, unclear intentions can lead supporters to employ inappropriate strategies.<n>We propose the Intention-centered Emotional Support Conversation framework.<n>It defines the possible intentions of supporters, identifies key emotional state aspects for inferring these intentions, and maps them to appropriate support strategies.
  arXiv  Detail & Related papers  (2025-06-06T10:14:49Z)
• R-CAGE: A Structural Model for Emotion Output Design in Human-AI Interaction [2.1828601975620257]
  R-CAGE is a theoretical framework for restructuring emotional output in long-term human-AI interaction.<n>By structurally regulating emotional rhythm, sensory intensity, and interpretive affordances, R-CAGE frames emotion not as performative output but as sustainable design unit.
  arXiv  Detail & Related papers  (2025-05-11T15:30:23Z)
• Towards Empathetic Conversational Recommender Systems [77.53167131692]
  We propose an empathetic conversational recommender (ECR) framework. ECR contains two main modules: emotion-aware item recommendation and emotion-aligned response generation. Our experiments on the ReDial dataset validate the efficacy of our framework in enhancing recommendation accuracy and improving user satisfaction.
  arXiv  Detail & Related papers  (2024-08-30T15:43:07Z)
• APTNESS: Incorporating Appraisal Theory and Emotion Support Strategies for Empathetic Response Generation [71.26755736617478]
  Empathetic response generation is designed to comprehend the emotions of others. We develop a framework that combines retrieval augmentation and emotional support strategy integration. Our framework can enhance the empathy ability of LLMs from both cognitive and affective empathy perspectives.
  arXiv  Detail & Related papers  (2024-07-23T02:23:37Z)
• K-ESConv: Knowledge Injection for Emotional Support Dialogue Systems via Prompt Learning [83.19215082550163]
  We propose K-ESConv, a novel prompt learning based knowledge injection method for emotional support dialogue system. We evaluate our model on an emotional support dataset ESConv, where the model retrieves and incorporates knowledge from external professional emotional Q&A forum.
  arXiv  Detail & Related papers  (2023-12-16T08:10:10Z)

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.