Titans-as-a-Layer: Test-Time Memory for Conversational Speech Emotion Recognition

• URL: http://arxiv.org/abs/2606.08573v1
• Date: Sun, 07 Jun 2026 11:07:30 GMT
• Title: Titans-as-a-Layer: Test-Time Memory for Conversational Speech Emotion Recognition
• Abstract summary: We study whether test-time neural memory can supply this missing context while leaving the large audio language models backbone intact.<n>We introduce a plug-and-play Memory-as-a-Layer (MAL) adapter that writes dialogue history into a small neural memory and reads it back as an audio-token-aligned residual update.
• Score: 11.240427393015034
• License: http://creativecommons.org/licenses/by-nc-nd/4.0/
• Abstract: Speech emotion recognition (SER) is commonly formulated as utterance-level classification, although conversational emotion depends on a speaker's usual vocal range and the emotional context established by previous utterances. Speech-language models provide strong pretrained acoustic and semantic representations, and can adapts them to SER labels via finetune, but this mechanism still missing per-dialogue state. We study whether test-time neural memory can supply this missing context while leaving the large audio language models (LALMs) backbone intact. Building on Titans, we introduce a plug-and-play Memory-as-a-Layer (MAL) adapter that writes dialogue history into a small neural memory and reads it back as an audio-token-aligned residual update, avoiding changes to the host model's token positions. Across different audio LLMs and emotion recognition datasets evaluations, our design improves SER performs across different evaluation metrics, supporting test-time memory as a residual contextual mechanism for conversational SER.

Related papers

• Acoustic Cue Alignment in Audio Language Models for Speech Emotion Recognition [58.25449304752214]
  We study whether explicit acoustic cues are grounded when the raw audio is already available.<n>We deriving six interpretable acoustic concept tokens from the standardised eGeMAPS paralinguistic feature set.<n>Aligned tokens improve unweighted average recall (UAR), whereas shuffled, conflicting, or corrupted tokens reduce performance.<n>We argue that token-only interventions provide a practical way to probe audio-grounded cue use, robustness, and interpretability in ALM-based affective computing.
  arXiv  Detail & Related papers  (2026-06-05T14:26:06Z)
• Speech Emotion Recognition Using MFCC Features and LSTM-Based Deep Learning Model [0.0]
  This study presents a speech emotion recognition system using Mel-Frequency Cepstral Coefficients (MFCCs) as features and a deep learning LSTM neural network.<n>A Support Vector Machine (SVM) with an RBF kernel served as a classical baseline, achieving 98% accuracy, against which the proposed LSTM model, achieving 99% accuracy, was validated.
  arXiv  Detail & Related papers  (2026-04-16T06:53:04Z)
• Detecting Mental Manipulation in Speech via Synthetic Multi-Speaker Dialogue [12.181747090385612]
  Mental manipulation is the strategic use of language to covertly influence or exploit others.<n>We present the first study of mental manipulation detection in spoken dialogues.<n>Using few-shot large audio-language models and human annotation, we evaluate how modality affects detection accuracy and perception.
  arXiv  Detail & Related papers  (2026-01-13T09:02:08Z)
• Lost in Phonation: Voice Quality Variation as an Evaluation Dimension for Speech Foundation Models [22.710371114925763]
  Speech foundation models (SFMs) have enabled the direct processing of spoken language from raw audio, bypassing intermediate textual representations.<n>This capability allows SFMs to be exposed to, and potentially respond to, rich paralinguistic variations embedded in the input speech signal.<n>We introduce a new parallel dataset featuring synthesized modifications to voice quality, designed to evaluate SFM responses to creaky and breathy voice.
  arXiv  Detail & Related papers  (2025-10-29T14:44:44Z)
• Moshi: a speech-text foundation model for real-time dialogue [78.88479749811376]
  Current systems for spoken dialogue rely on pipelines independent voice activity detection and text-to-speech. We show how Moshi Moshi can provide streaming speech recognition and text-to-speech. Our resulting model is first real-time full spoken large language model modality.
  arXiv  Detail & Related papers  (2024-09-17T17:55:39Z)
• Paralinguistics-Enhanced Large Language Modeling of Spoken Dialogue [71.15186328127409]
  Paralinguistics-enhanced Generative Pretrained Transformer (ParalinGPT) Model takes the conversational context of text, speech embeddings, and paralinguistic attributes as input prompts within a serialized multitasking framework. We utilize the Switchboard-1 corpus, including its sentiment labels as the paralinguistic attribute, as our spoken dialogue dataset.
  arXiv  Detail & Related papers  (2023-12-23T18:14:56Z)
• Inner speech recognition through electroencephalographic signals [2.578242050187029]
  This work focuses on inner speech recognition starting from EEG signals. The decoding of the EEG into text should be understood as the classification of a limited number of words (commands) Speech-related BCIs provide effective vocal communication strategies for controlling devices through speech commands interpreted from brain signals.
  arXiv  Detail & Related papers  (2022-10-11T08:29:12Z)
• SpeechLM: Enhanced Speech Pre-Training with Unpaired Textual Data [100.46303484627045]
  We propose a cross-modal Speech and Language Model (SpeechLM) to align speech and text pre-training with a pre-defined unified representation. Specifically, we introduce two alternative discrete tokenizers to bridge the speech and text modalities. We evaluate SpeechLM on various spoken language processing tasks including speech recognition, speech translation, and universal representation evaluation framework SUPERB.
  arXiv  Detail & Related papers  (2022-09-30T09:12:10Z)
• Textless Speech Emotion Conversion using Decomposed and Discrete Representations [49.55101900501656]
  We decompose speech into discrete and disentangled learned representations, consisting of content units, F0, speaker, and emotion. First, we modify the speech content by translating the content units to a target emotion, and then predict the prosodic features based on these units. Finally, the speech waveform is generated by feeding the predicted representations into a neural vocoder.
  arXiv  Detail & Related papers  (2021-11-14T18:16:42Z)
• CoMPM: Context Modeling with Speaker's Pre-trained Memory Tracking for Emotion Recognition in Conversation [0.0]
  We introduce a context embedding module (CoM) combined with a pre-trained memory module (PM) We show that the pre-trained memory significantly improves the final accuracy of emotion recognition. We experimented on both the multi-party datasets (MELD, EmoryNLP) and the dyadic-party datasets (IEMOCAP, DailyDialog)
  arXiv  Detail & Related papers  (2021-08-26T07:45:09Z)
• Seen and Unseen emotional style transfer for voice conversion with a new emotional speech dataset [84.53659233967225]
  Emotional voice conversion aims to transform emotional prosody in speech while preserving the linguistic content and speaker identity. We propose a novel framework based on variational auto-encoding Wasserstein generative adversarial network (VAW-GAN) We show that the proposed framework achieves remarkable performance by consistently outperforming the baseline framework.
  arXiv  Detail & Related papers  (2020-10-28T07:16:18Z)

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.