PulseAugur
LIVE 08:13:16
research · [14 sources] ·
0
research

AI models struggle with emotion nuance, researchers explore new evaluation and generation methods

Researchers are exploring the nuances of emotion in AI, with several papers focusing on Large Language Models (LLMs) and speech processing. One study investigates how well small language models preserve emotions during machine translation across several European languages. Another paper introduces a new dataset and pipeline for speech captioning that accounts for emotion transitions in discourse. Additionally, research critically examines the metrics used to evaluate emotional expressiveness in speech generation, questioning the reliance on embedding similarity. Finally, a study analyzes how LLMs infer emotions, identifying internal mechanisms and proposing methods to improve their emotion recognition capabilities, while also highlighting the gap between LLM annotations and human judgment. AI

Summary written by gemini-2.5-flash-lite from 14 sources. How we write summaries →

IMPACT Advances in understanding and generating emotional AI could lead to more nuanced human-AI interactions and improved affective computing applications.

RANK_REASON Multiple academic papers published on arXiv exploring various aspects of emotion in AI systems.

Read on arXiv cs.CV →

COVERAGE [14]

  1. arXiv cs.CL TIER_1 · Keito Inoshita, Xiaokang Zhou, Akira Kawai, Katsutoshi Yada ·

    LLMs Capture Emotion Labels, Not Emotion Uncertainty: Distributional Analysis and Calibration of Human--LLM Judgment Gaps

    arXiv:2604.27345v1 Announce Type: new Abstract: Human annotators frequently disagree on emotion labels, yet most evaluations of Large Language Model (LLM) emotion annotation collapse these judgments into a single gold standard, discarding the distributional information that disag…

  2. arXiv cs.AI TIER_1 · Dawid Wisniewski, Igor Czudy ·

    Beyond Semantics: Measuring Fine-Grained Emotion Preservation in Small Language Model-Based Machine Translation

    arXiv:2604.27920v1 Announce Type: cross Abstract: Preserving affective nuance remains a challenge in Machine Translation (MT), where semantic equivalence often takes precedence over emotional fidelity. This paper evaluates the performance of three state-of-the-art Small Language …

  3. arXiv cs.CL TIER_1 · Igor Czudy ·

    Beyond Semantics: Measuring Fine-Grained Emotion Preservation in Small Language Model-Based Machine Translation

    Preserving affective nuance remains a challenge in Machine Translation (MT), where semantic equivalence often takes precedence over emotional fidelity. This paper evaluates the performance of three state-of-the-art Small Language Models (SLMs) -- EuroLLM, Aya Expanse, and Gemma -…

  4. arXiv cs.CL TIER_1 · Yun-Shao Tsai, Yi-Cheng Lin, Huang-Cheng Chou, Tzu-Wen Hsu, Yun-Man Hsu, Chun Wei Chen, Shrikanth Narayanan, Hung-yi Lee ·

    The False Resonance: A Critical Examination of Emotion Embedding Similarity for Speech Generation Evaluation

    arXiv:2604.26347v1 Announce Type: cross Abstract: Objective metrics for emotional expressiveness are vital for speech generation, particularly in expressive synthesis and voice conversion requiring emotional prosody transfer. To quantify this, the field widely relies on emotion s…

  5. arXiv cs.CL TIER_1 · Shuhao Xu, Yifan Hu, Jingjing Wu, Zhihao Du, Zheng Lian, Rui Liu ·

    EmoTransCap: Dataset and Pipeline for Emotion Transition-Aware Speech Captioning in Discourses

    arXiv:2604.26417v1 Announce Type: new Abstract: Emotion perception and adaptive expression are fundamental capabilities in human-agent interaction. While recent advances in speech emotion captioning (SEC) have improved fine-grained emotional modeling, existing systems remain limi…

  6. arXiv cs.CL TIER_1 · Katsutoshi Yada ·

    LLMs Capture Emotion Labels, Not Emotion Uncertainty: Distributional Analysis and Calibration of Human--LLM Judgment Gaps

    Human annotators frequently disagree on emotion labels, yet most evaluations of Large Language Model (LLM) emotion annotation collapse these judgments into a single gold standard, discarding the distributional information that disagreement encodes. We ask whether LLMs capture the…

  7. arXiv cs.CL TIER_1 · Rui Liu ·

    EmoTransCap: Dataset and Pipeline for Emotion Transition-Aware Speech Captioning in Discourses

    Emotion perception and adaptive expression are fundamental capabilities in human-agent interaction. While recent advances in speech emotion captioning (SEC) have improved fine-grained emotional modeling, existing systems remain limited to static, single-emotion characterization w…

  8. arXiv cs.CL TIER_1 · Hung-yi Lee ·

    The False Resonance: A Critical Examination of Emotion Embedding Similarity for Speech Generation Evaluation

    Objective metrics for emotional expressiveness are vital for speech generation, particularly in expressive synthesis and voice conversion requiring emotional prosody transfer. To quantify this, the field widely relies on emotion similarity between reference and generated samples.…

  9. arXiv cs.CL TIER_1 · Bangzhao Shu, Arinjay Singh, Mai ElSherief ·

    From Syntax to Emotion: A Mechanistic Analysis of Emotion Inference in LLMs

    arXiv:2604.25866v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly used in emotionally sensitive human-AI applications, yet little is known about how emotion recognition is internally represented. In this work, we investigate the internal mechanisms of …

  10. arXiv cs.CL TIER_1 · Taryn Wong, Zeerak Talat, Hanan Aldarmaki, Anjalie Field ·

    Unrequited Emotions: Investigating the Gaps in Motivation and Practice in Speech Emotion Recognition Research

    arXiv:2604.25776v1 Announce Type: new Abstract: Critical analyses of emotion recognition technology have raised ethical concerns around task validity and potential downstream impacts, urging researchers to ensure alignment between their stated motivations and practice. However, t…

  11. arXiv cs.CL TIER_1 · Mai ElSherief ·

    From Syntax to Emotion: A Mechanistic Analysis of Emotion Inference in LLMs

    Large language models (LLMs) are increasingly used in emotionally sensitive human-AI applications, yet little is known about how emotion recognition is internally represented. In this work, we investigate the internal mechanisms of emotion recognition in LLMs using sparse autoenc…

  12. arXiv cs.CL TIER_1 · Anjalie Field ·

    Unrequited Emotions: Investigating the Gaps in Motivation and Practice in Speech Emotion Recognition Research

    Critical analyses of emotion recognition technology have raised ethical concerns around task validity and potential downstream impacts, urging researchers to ensure alignment between their stated motivations and practice. However, these discussions have not adequately influenced …

  13. arXiv cs.CL TIER_1 · He Hu, Lianzhong You, Hongbo Xu, Qianning Wang, Fei Richard Yu, Fei Ma, Zebang Cheng, Zheng Lian, Yucheng Zhou, Laizhong Cui ·

    EmoBench-M: Benchmarking Emotional Intelligence for Multimodal Large Language Models

    arXiv:2502.04424v4 Announce Type: replace Abstract: With the integration of multimodal large language models (MLLMs) into robotic systems and AI applications, embedding emotional intelligence (EI) capabilities is essential for enabling these models to perceive, interpret, and res…

  14. arXiv cs.CV TIER_1 · He Hu, Tengjin Weng, Zebang Cheng, Yu Wang, Jiachen Luo, Bj\"orn Schuller, Zheng Lian, Laizhong Cui ·

    EmoTrans: A Benchmark for Understanding, Reasoning, and Predicting Emotion Transitions in Multimodal LLMs

    arXiv:2604.23348v1 Announce Type: new Abstract: Recent multimodal large language models (MLLMs) have shown strong capabilities in perception, reasoning, and generation, and are increasingly used in applications such as social robots and human-computer interaction, where understan…