arXiv_SD
-
Self-Localization of Distributed Microphone Arrays Using One Source and Acoustic Reflections
2026-10-30 10:43:55Thomas Guffroy (MACARON, UMRAE), Jean-Daniel Pascal Prieto (UMRAE, MACARON), C\'edric Foy (UMRAE), Antoine Deleforge (MACARON)Abstract
We present a method to estimate the relative position and orientation of two non-coplanar microphone arrays using a room impulse response (RIR) from a single source. The approach exploits early acoustic reflections, retrieved by each array in the form of a 3D image-source point cloud, using a recently-proposed image-source localization technique. The task is then reduced to aligning the two unlabeled point clouds, which is made difficult by their inherent symmetries and by spurious, missing and noisy estimates. To this aim we introduce ARCTIC, a point-cloud registration algorithm that performs a ranked search over congruent triangles anchored on the true source. Simulated and real experiments show that ARCTIC strongly outperforms two baselines, achieving mean angular and radial self-localization errors below 5 and 10 cm across 36 real RIRs measured in 6 different rooms.
Abstract (translated)
URL
https://arxiv.org/abs/2610.41547
PDF
-
ARPEGE: Dataset and Methods to Estimate Acoustic and Geometric Parameters of Rooms from Sound
2026-10-30 10:41:28Jean-Daniel Pascal Prieto (UMRAE, MACARON), Prerak Srivastava (UMRAE), C\'edric Foy (UMRAE), Antoine Deleforge (MACARON)Abstract
Can one hear the walls of a room? To foster audio signal processing research around this question we release ARPEGE, a dataset containing hundreds of ambisonic spherical-array room impulse responses recorded in 6 empty rooms, annotated by the relative positions of sources, microphones and all acoustic reflectors. The dataset includes multiple room sizes, surface materials, orientations and distances, as well as recordings of human speakers and handclaps. Two state-of-the-art methods on image source localization and blind room parameter estimation are demonstrated on the data. Initial code for re-simulating the measured room impulse responses is provided, a longer term goal being to digitally twin the dataset.
Abstract (translated)
URL
https://arxiv.org/abs/2610.41543
PDF
-
Scalable Audio-LLM System for Speech MOS Evaluation: A Diagnostic Case Study and Adaptation Roadmap
2026-10-30 10:37:06Dimitrios Kokkinis, Panagiotis Kakoulidis, Nikolaos Ellinas, Chrysovalanti Lymperi, Alexandra Vioni, Georgios Vardaxoglou, Iakovi Alexiou, Myrsini Christidou, Jeasung Lee, Gunu Jho, Inchul Hwang, Pirros Tsiakoulis, Aimilios ChalamandarisAbstract
The continuous deployment of enterprise voice systems creates a critical operational bottleneck: evaluating the acoustic quality of synthetic speech across high-volume production updates. While human Mean Opinion Score (MOS) remains the quality gold standard, it introduces severe operational costs and turn-around delays that make continuous integration infeasible. In this work, we present a system design case study for a scalable, low-cost and offline evaluation pipeline capable of scoring large batches of synthetic speech at marginal cost. First, we use Qwen2-Audio-7B-Instruct in a zero-shot setting for granular MOS estimation, noticing perceptual blind spots and inherent biases. Then, we analyze SOMOS dataset and construct a diagnostic benchmark that operationalizes two prosodic dimensions associated with human naturalness ratings. Finally, we adapt Qwen2-Audio-7B-Instruct through a parameter-efficient finetuning (LoRA) on the task of MOS evaluation, observing improvements over base model and widely used neural baselines like UTMOS and establishing a practical blueprint for enterprise-grade speech quality assurance.
Abstract (translated)
URL
https://arxiv.org/abs/2610.41539
PDF
-
Hierarchical Dual-Graph Retrieval for Long-form Audio Reasoning
Abstract
Large audio language models (LALMs) struggle with long-form audio reasoning, as processing entire recordings often limits to capture query-relevant content distributed across distant segments while incurring substantial computational cost. We introduce AudioHelix, a hierarchical DualGraph framework tailored for LALMs' retrieval-augmented generation (RAG) systems. Unlike existing text-only or non-graph RAG approaches, AudioHelix connects related textual and acoustic evidence across temporally distant segments for multi-hop reasoning. Extensive experiments on TraceAV-Bench, MMAU-Pro, and AudioMarathon demonstrate that AudioHelix consistently outperforms full-audio processing and existing RAG methods. The project site is this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/2610.41390
PDF
-
From prediction error to sparse events: Biologically inspired signed-dyadic coding for Mandarin keyword spotting
2026-10-30 07:46:45Zijian Zhou, Rui Liu, Rongxiang Nie, Rui Ma, Ziyao Zhang, Xin DuAbstract
In continuous spiking keyword spotting, temporally correlated acoustic frames can repeatedly trigger memory access and neural computation at a dense input interface, even when hidden activity is sparse. Predictive processing and sensory adaptation suggest retaining predictable input locally and responding selectively to deviations. Inspired by this mechanism, we propose Predictive Signed-Dyadic Coding (PG). PG maintains a local prediction for each input channel and transmits an event only when the residual exceeds a learned threshold. Each event carries a sign and a finite power-of-two magnitude, and the receiver reconstructs it with fixed-point shifts and additions. In Mandarin keyword spotting, INPUT-PG reaches 90.00% accuracy while reducing mean SRAM reads by 33.87% and inference cycles by 32.77%. PG provides a hardware interface for research on low-power continuous speech processing and event-based sensory systems.
Abstract (translated)
URL
https://arxiv.org/abs/2610.41381
PDF
-
TAPE: Learning General Audio Effect Representations
2026-10-29 17:17:15Phillip Long, Yongyi Zang, Qihui Yang, Christodoulos Benetatos, Junfei Liu, Randal Leistikow, Julian McAuley, Perry R. CookAbstract
Musicians and mixing engineers often can decipher the signal processing chain they hear on a reference recording, then adapt it to new material. We present TAPE (Transformations of Audio Parameter Embeddings), which estimates this processing from the processed reference audio alone. An encoder trained to identify processors in eight major effect families maps the reference to an embedding. A differentiable digital signal processing (DSP) chain decodes the embedding into editable parameters in one forward pass and applies them to a different unprocessed recording. Its modules are equalization, compression, waveshaping, a small residual network, and reverberation. TAPE achieves lower spectral error than DeepAFx-ST, the previous state of the art for single-step regression, and lower spectral error than per-example search over a chain approximating ST-ITO's, while running 275$\times$ faster than that search. Because the decoder outputs DSP parameters, its chain can also run causally: a C++ implementation with the trained weights processed 128-sample buffers 1.8$\times$ faster than real time, with no missed deadlines.
Abstract (translated)
URL
https://arxiv.org/abs/2610.40759
PDF
-
JYVA at BeTraC: Audio-Based Medical Note Generation with SOAP Section-Specific Queries
2026-10-28 10:22:57Youness Dkhissi (LIUM), Valentin Vielzeuf (IRT b-com), Antoine Caubri\`ere (IRT b-com), Jean-Emile Dartois (IRT b-com)Abstract
We present our submission to the Beyond Transcription Challenge (BeTraC) on medical note generation from doctor-patient audio conversations. We propose a section-aware approach based on Qwen2.5-Omni-3B, in which a lightweight Q-Former learns SOAP section-specific queries to extract compact clinical representations from speech. In the lightweight track, our system ranks fourth and substantially outperforms the challenge baseline, achieving a Concept F1 of 0.5005 on the official test set. The organizers' analysis and our heuristic evaluation suggest improved robustness to hallucinations, and better stability under difficult audio conditions. These results show that explicitly modeling SOAP structure through section-specific queries is an effective way to improve the faithfulness and reliability of audio-based medical note generation.
Abstract (translated)
URL
https://arxiv.org/abs/2610.40695
PDF
-
Partially Edited Audio: Generation, Defense, and Open Problems
2026-10-27 17:54:42Lin Zhang, You Zhang, Nicholas Evans, Xin Wang, David Harwath, Abhinav Dhall, Bowen Shi, Chengzhe Sun, Siwei Lyu, Sanjeev KhudanpurAbstract
Recent advances in generative audio models facilitate increasingly flexible and selective editing of linguistic content and paralinguistic, extralinguistic, and other attributes. This paradigm, named partially edited audio, differs from full audio generation as only portions of the real recording are edited. This technique can serve both beneficial and harmful applications. We therefore present a concise survey from both generation and defense perspectives, focusing on audio, as a contribution to the ICASSP 2027 special session entitled "Partially Edited Media: Perspectives from Synthesis and Defense." We first define partially edited audio and argue that what counts as ``fake'' or harmful is an application-dependent judgment, not a fixed property of the edit. We then survey audio editing by editing target, approaches, and evaluation, and then defense methods by detection, localization, diarization, and interpretability, summarizing representative benchmarks and toolkits. Finally, we discuss future directions and open problems.
Abstract (translated)
URL
https://arxiv.org/abs/2610.37640
PDF
-
Beyond Answer Changes: Matched Same-Class Controls for Local Audio Evidence
2026-10-27 17:36:06Yunuo Wang, Maoyuan Chen, Yuanjin Huang, Xing LiuAbstract
For audio-language models, an answer change after a local sound-event edit can reflect a response to the event's class, sensitivity to waveform rewriting, or both. We compare matched same-class and different-class replacements of a queried event. Both use the same interval, level estimate, donor fitting, and crossfade, while the same-class replacement preserves the correct answer. Across 600 synthetic soundscapes and three models from two architecture families, different-class replacements produce 49.7 to 60.9 percentage points more answer changes than same-class replacements. Same-class accuracy remains within 5.2 percentage points of original accuracy. Among trials with correct original and same-class answers, soundscape-weighted selective accuracy for the inserted class is 73.7 to 81.9%, averaging eligible trials within each soundscape and then weighting eligible soundscapes equally. This protocol tests how model answers respond to local changes in event class.
Abstract (translated)
URL
https://arxiv.org/abs/2610.37605
PDF
-
EmbeddingGemma 2: Unified and Efficient Multimodal Representations
2026-10-27 15:00:42Henrique Schechter Vera, Sahil Dua, Jovan Andonov, Sara Smoot, Sebastian Russo, Ivan Llanos, Max Matveev, Jae Hyeon Yoo, Kat Black, Shih-Cheng Huang, Josiah Aklilu, Maarten Grootendorst, Omar Sanseviero, Chandramouli Amarnath, Ian Ballantyne, Zhe Li, Daniel Salz, Ibrahim Alabdulmohsin, Ilya Philippov, Kristen Quan, Glenn Cameron, Edouard Yvinec, Luciano Martins, Doug Reid, Sindhu Raghuram Panyam, Weiyi Wang, Hengxuan Ying, Madhuri Shanbhogue, Tanmaya Dabral, Shanfeng Zhang, Juyeong Ji, Renjie Wu, Joe Zou, Xu Chen, Yunan Zhang, Yu-Hui Chen, Matthew Soulanille, Vinod Mamillapalli, Yi-Chun Kuo, Wai Hon Law, Matt Kreileder, Alice Zheng, Suril Shah, Mogan Shieh, Pooja Consul, Alberto Montes, Andreas Hess, Andres Sevilla, Babak Samari, Brendan McCloskey, Brian Potetz, Daniel Cer, Georgios Kathareios, Gustavo Hern\'andez \'Abrego, Hesen Zhang, Ji Ma, Jiaxing Wu, Keegan Mosley, Paul Suganthan, Rui Meng, Shen Gao, Shengcao Cao, Yi-Ting Chen, Yichang Chen, Zhongli Ding, Kendra Byrne, Cormac Brick, Olivier Lacombe, Qin Yin, Tris Warkentin, Armand Joulin, Tom Duerig, Mojtaba SeyedhosseiniAbstract
We present EmbeddingGemma 2, an open multimodal embedding model which natively represents text, images, audio, video, and combinations thereof in a unified space. Our approach introduces improvements such as simplified per-layer embeddings and a multi-stage curriculum with modality and objective progression which ensure strong quality across unimodal, cross-modal, and mixed-modal evaluations. Notably, in comprehensive embedding benchmarks across all modalities, EmbeddingGemma 2 establishes a new Pareto frontier, outperforming both specialist models and models two to twelve times larger. Furthermore, our design offers flexibility for resource-constrained environments, supporting selective modality encoder loading, adjustable vision token budgets, truncatable embedding dimensions, and quantized weights. We present ablations analyzing key decisions and release EmbeddingGemma 2 to promote further research.
Abstract (translated)
URL
https://arxiv.org/abs/2610.37364
PDF
-
Understanding Everyday Communication Patterns, Stress, and Productivity in Hybrid Work through Egocentric Wearable Audio Sensing
2026-10-27 14:02:16Tiantian Feng, Aditya Kommineni, Emily Zhou, Burcin Becerik-Gerber, Gale Lucas, Shawn C. Roll, Shrikanth NarayananAbstract
The hybrid work trend in the post-COVID period has transformed how people communicate while at work. However, there is a lack of quantitative understanding of how home and office environments shape communicative behavior and affective well-being. To study this, as a part of a larger study of the role of work environment on stress, we designed an egocentric wearable audio sensing system that captured interpersonal communication across both working locations. The system extracts real-time acoustic features on a wearable audio sensor and performs offline foreground speaker diarization to derive speaking load and activeness features. To improve diarization performance under low-dimensional feature constraints, we apply a teacher-student knowledge distillation framework that transfers knowledge from large-scale speech foundation models into a compact ResNet-based model, achieving consistent improvements in diarization performance over the baseline across three public egocentric speech datasets. In a four-month longitudinal deployment with 16 participants, our system revealed location-dependent differences in work communication patterns. Moreover, the mixed linear model analyses showed that office-based speaking load was significantly associated with higher self-reported stress and lower self-reported productivity, while speaking activeness during home-based work amplified both productivity and affective load in this cohort.
Abstract (translated)
URL
https://arxiv.org/abs/2610.37270
PDF
-
EEG-Based Decoding of Musical Mode and Melodic Direction Using Empirical Mode Decomposition
2026-10-27 12:47:51Sergio Guill\'en, Lorenzo J. Tard\'on, Ana M. Barbancho, Isabel BarbanchoAbstract
Electroencephalography (EEG) has proven to be a powerful tool for decoding neural activity in real-life scenarios, including the perception and processing of emotions elicited by music. Among the various musical features that modulate these responses, musical mode and melodic direction (or contour) play a central role. This work proposes a novel EEG feature extraction framework based on Empirical Mode Decomposition (EMD) to decompose EEG signals recorded during active piano performance for the detection of these two musical features. Energy-based descriptors were computed from the first four Intrinsic Mode Functions (IMFs) obtained through EMD and used to train a Wide Neural Network classifier. The model performance was benchmarked against several standard classifiers commonly employed in EEG classification tasks. Two experimental scenarios were explored: intra-subject and intersubject. In the former case, the proposed framework achieved average accuracies of 84%, whereas in the inter-subject case, accuracies reached 81% on average using the first IMF, with AUC-ROC values of 0.9 on average in both cases. A frequency-domain analysis revealed a strong correspondence between the IMFs and conventional EEG sub-bands, except for certain specific frequency components that were not fully represented. Predominant activations were observed over the frontal, temporal, and parietooccipital regions, areas consistently associated with music processing. Furthermore, the proposed method demonstrated robustness to handedness, successfully classifying EEG data across participants regardless of their dominant hand. These findings highlight the potential of EMD-based feature extraction for identifying music-related neural patterns and contribute to advancing research in Music Information Retrieval (MIR), affective neuroscience, and the broader understanding of the neural encoding of musical structure.
Abstract (translated)
URL
https://arxiv.org/abs/2610.37148
PDF
-
Design and Architecture of Airmod: An Active Noise Cancellation and Deep Learning System for Real-Time Respiratory Rate Estimation from Tracheal Sounds
2026-10-27 11:46:14Shang-Ran Huang, Fu-Shun Hsu, Ji-Chang Huang, Chuan Chen, Shou-Zen Fan, Chang-Fu Su, Eng-Che Huang, Yuan-Ren ChengAbstract
Tracheal auscultation is recommended by clinical practice guidelines, and required by some U.S. state regulations, for monitoring ventilation during procedural sedation and monitored anesthesia care. However, continuous auscultation with a conventional stethoscope is impractical, susceptible to ambient noise, and difficult to interpret. We developed the Artificial Intelligence Respiratory MOnitoring Device (Airmod), an FDA-cleared software as a medical device (510(k) K242798) that continuously monitors tracheal sounds and estimates breath-to-breath respiratory rate in real time. In this paper, we describe its software architecture and key processing modules. Ambient noise is suppressed by active noise control, using a normalized least-mean-square adaptive filter with a dual-microphone electronic stethoscope. A multiclass convolutional bidirectional gated recurrent unit (CNN-BiGRU) model detects inhalations and exhalations in 15-s inference windows. It was trained on HF_Tracheal_V2, a new database of 373 continuous tracheal sound recordings (8,521 min; 107,644 labeled inhalations) from 369 subjects under sedation. To handle continuous acoustic streaming, Airmod runs inference on overlapping 15-s windows. It refines the detected inhalations by removing incomplete events, low-energy false positives, and double counts, concatenates them into a history list, and estimates respiratory rate from the mean period of the last three breaths. On the test set, at the default 2-s prediction interval with refinement enabled, inhalation detection achieved a sensitivity of 0.828, a positive predictive value of 0.832, and an F1-score of 0.830. Airmod also broadcasts the noise-cancelled tracheal sound and displays a real-time spectrogram to help clinicians recognize respiratory abnormalities.
Abstract (translated)
URL
https://arxiv.org/abs/2610.37083
PDF
-
Perceptual Differences and Preferences in Speech Enhancement
2026-10-27 09:27:02Simon Dahl Jepsen, Jesper Rindom JensenAbstract
Speech enhancement (SE) systems are commonly evaluated with objective metrics, with and without references, and with categorical subjective listening tests, all measuring speech quality and intelligibility. So far, it is unclear from such evaluations whether the processed speech is perceptually different or is preferred by everyday listeners. We investigate discrimination and preference as complementary dimensions of SE evaluation through crowdsourced triangle and pairwise preference tests across four SE systems, with comparisons to clean references. Results show that both listeners' ability to distinguish and whether they exhibit systematic preference vary across systems and substantially across acoustic conditions. Pairwise evaluations of objective metrics showed variable correlation with both discriminability and preference, motivating further studies into these novel aspects in speech enhancement evaluation.
Abstract (translated)
URL
https://arxiv.org/abs/2610.36892
PDF
-
Physiology-Informed Audio Quality and Listening Effort Estimation
Abstract
Audio systems aim to maximise the listener's perceived quality, traditionally evaluated through mean opinion scoring (MOS), a subjective process that is slow, costly, and collected only after playback. Machine learning now predicts MOS directly from the audio, but in generalising across participants it tends to discard the subject variance that defines individual perception. Neurophysiological measures may offer a complementary signal taken directly from the listener, objective, implicit, and continuous, though prior work has relied on laboratory-grade EEG and fNIRS. We ask how practical consumer-grade devices compare, and whether they might recover population-level audio-quality trends and, when calibrated per listener, track that listener's effort. Recording 32-channel laboratory EEG alongside wearable EEG, fNIRS, and eye tracking from thirty listeners, our results suggest commercial wearable and optical sensing can decode audio quality at statistically significant levels and recover listening effort per listener. These findings may point towards platforms tuned not only to perceived quality but to the effort a clip demands, read from the listener's own response as they listen.
Abstract (translated)
URL
https://arxiv.org/abs/2610.36865
PDF
-
Expressive Guitar Synthesis via Controllable Latent Flow Matching
Abstract
Realistic guitar synthesis requires both audio fidelity and interpretable control over how notes are performed. The input timed MIDI events decide what is played but not how, leaving the continuous within-note decisions of guitar performance--such as pitch inflection and per-string articulation--unspecified. A model that directly maps MIDI to audio must therefore realize the performance implicitly, and users are left with limited control over the synthesized performance. We therefore propose a two-stage guitar synthesis pipeline that predicts editable, per-string $f_0$ and amplitude trajectories from timed note events and note-level dynamics, and renders them into audio using latent flow matching. Experimental results show that the proposed system outperforms a matched single-stage counterpart and DDSP-Guitar in control fidelity and perceived quality, while keeping the controls editable. Sound examples can be found online.
Abstract (translated)
URL
https://arxiv.org/abs/2610.36786
PDF
-
S3-S2ST: Semantically Supervised Streaming Speech-to-Speech Translation
2026-10-27 07:44:28Yuxiang Zhao, Zhiyun Fan, Yu Pu, Yanjie An, Yushen Chen, Wenrui Liang, Baiji Liu, Yunchong Xiao, Zhaoci Liu, Keqi Deng, Kai Yu, Zhifu Gao, Xiangang Li, Xie ChenAbstract
Streaming speech-to-speech translation requires timely decisions about when partial input is sufficient for accurate translation. We present S3-S2ST, a model that learns to translate and speak from semantically aligned streaming supervision. By converting utterance-level speech pairs into incremental training examples, we teach a learned policy when to commit translations. Its adjustable threshold controls the quality--latency tradeoff on a fixed 640\,ms input grid. S3-S2ST delivers competitive English--Chinese translation on FLEURS and leads the evaluated streaming systems in translation quality on the internal 9-pair test set. On RealSI, adapting the threshold to buffered speech further improves quality over a low fixed threshold with little change in first-audio latency. We plan to release the model and code.
Abstract (translated)
URL
https://arxiv.org/abs/2610.36777
PDF
-
The Content Leash: Mitigating Degenerate Solutions in Invertible Speaker Anonymization
2026-10-27 07:28:59Nirjon Mobarak Talukder, Rafid Ahmed, Mohammad Ishrak Abedin, Abu Raihan Mostofa KamalAbstract
Voice anonymization aims to protect a speaker's identity while preserving the linguistic content of speech, enabling voice data to be shared and processed with reduced privacy risk. This work focuses on reversible voice anonymization, where the original voice can be recovered by an authorized user rather than being permanently discarded. However, existing approaches can achieve apparent speaker anonymization by degrading the speech itself, while their effectiveness may also depend on the speaker-verification representation used during training. This creates a need for methods that protect speaker identity without sacrificing speech intelligibility and that remain effective against unseen speaker verifiers. We introduce content preserving training constraints to improve the privacy-utility trade-off by preventing anonymization from being achieved primarily through speech degradation. We further introduce a verifier-generalization strategy that incorporates a stronger speaker-verification representation during training, improving anonymization robustness beyond the verifier used in the original training setup. The resulting approach maintains intelligible speech, strengthens anonymization, and preserves reliable voice restoration, with evaluation on unseen speakers and held-out speaker verification models. On 38 unseen speakers, the approach raises the attacker's equal error rate from a measured floor of 2.8% to 26.3%, and to 27.9% against a held-out verifier, with 11.8% word error rate and 94.9% speaker similarity on key-based restoration. These improvements advance reversible voice anonymization toward practical privacy-preserving speech applications, where anonymized speech can be safely processed or shared while the original voice remains recoverable under authorized conditions.
Abstract (translated)
URL
https://arxiv.org/abs/2610.36768
PDF
-
SMEBENCH: Benchmarking Neural Music Editing with Synthetized MIDI Edits
2026-10-27 06:31:02Matthieu Cervera, Francesco Paissan, Micael Antunes, Mirco Ravanelli, Ricard Marxer, Cem SubakanAbstract
Neural music editing requires a system to perform a requested change while preserving the musical content that should remain unchanged. Existing evaluations often lack a paired target edit, making these requirements difficult to measure separately. We introduce SMEBench, a large-scale benchmark of controlled music edits constructed by modifying symbolic MIDI performances and rendering matched source and target audio. Each example contains a source, an instruction, a paired target, edit metadata, and complementary descriptions. The benchmark covers twelve instrumentation, articulation, dynamics, and pitch operations and includes a balanced 3,600-example test set together with a smaller real-audio complement derived from multitrack recordings. The paired targets allow source preservation, target agreement, and textual compatibility to be evaluated independently. We use these references to diagnose common audio-editing metrics and benchmark six representative editing systems, showing that the metric used to rank a system can materially change the conclusion. The dataset is available at our repository.
Abstract (translated)
URL
https://arxiv.org/abs/2610.36715
PDF
-
From Voiceprints to User Entities: Online Speaker-Memory Construction for Multi-User Spoken Agents
2026-10-27 06:27:24Yucong Zhang, Cancan Li, Bo Xiao, Ke Zeng, Ming LiAbstract
Personalized spoken agents must identify speakers before accessing user-specific memory, yet memory benchmarks typically assume correctly attributed histories. We argue that voiceprints are observations of identity, not identities themselves, and formulate online speaker-memory construction to maintain persistent user entities as acoustic evidence evolves. We introduce EvoSpeakerMem, a training-free layer for causal speaker routing, selective acoustic-state updates, and identity-partitioned memory writes and retrieval through a replaceable text-memory backend. We also construct LoCoMo-ID, a spoken-memory benchmark with hidden speaker identities, speaker-sensitive queries, simulated recording-condition shifts, and four-speaker identity-confusion tests. On longitudinal SMIIP-TV, genuine-probe rejection under a frozen near-enrollment threshold rises from 29.8% to 55.6% over approximately one year. On LoCoMo-ID speaker-sensitive queries, gated updating improves identity-gated QA over fixed representations by 3.86 and 11.62 percentage points under clean and shifted conditions, respectively, with no wrong-user authorizations. The advantage persists in the four-speaker test. Yet reducing fragmentation and improving QA can coexist with incorrect cross-user links, revealing a fragmentation--contamination trade-off that QA alone cannot capture.
Abstract (translated)
URL
https://arxiv.org/abs/2610.36713
PDF
