call center technology teams

Abstract: End-of-turn (EoT) detection has been studied mostly through features, models, and modalities, while the speech genre of the training data has drawn far less attention. We isolate this factor. Using one reproducible annotation pipeline, we build a matched pair of corpora, multi-turn human-human conversational and single-turn human-machine, and train the same streaming ConvLSTM EoT detector on each. Across four test sets the conversational model reaches up to +0.45 higher F1, driven by fewer false alarms on the mid-turn pauses and hesitations that are rare in single-turn corpora. This conversational-training effect is not specific to frame-level decoding. Under a matched complete-or-not test it also appears at the utterance level, across two encoders, but only on conversational test data. Testing only on single-turn speech, as most prior work does, therefore hides the benefit of conversational training data. Against open-source baselines, our 1.8M-parameter model attains the highest F1 on every test set.

Abstract: Automatic speech recognition (ASR) systems often perform less accurately for speakers of regional dialects, but less is known about how users respond to these limitations and how regional voices produced by voice assistants (VAs) are evaluated through a sociolinguistic lens. This study investigates these issues among speakers of Newcastle English (Geordies), examining both linguistic accommodation to voice technologies and perceptions of regional versus mainstream-accented VA voices. Participants interacted with a simulated VA that spoke either Newcastle English or Standard Southern British English (SSBE). They subsequently completed semi-structured interviews about their experiences with the experimental VA and voice technologies, language use, accent attitudes, and place orientation. The findings reveal a two-way process of sociolinguistic bias. The Geordie VA was associated with familiarity, warmth, personality, and local belonging, but also with concerns about professionalism and authority, whereas the SSBE voice was perceived as more neutral, clear, and credible. Participants also reported adapting their speech when voice technologies failed to recognise them, most commonly by shifting to a more mainstream accent, enunciating more clearly, and speaking more slowly. These strategies are similar to accommodation in human-human interaction and reflect users' perceptions of what technologies can recognise, also known as perceived intelligibility pressures. Overall, the findings demonstrate that voice technologies can reproduce existing accent biases both by encouraging users to suppress regional features and by activating established social stereotypes through synthetic voices. The study positions sociolinguistics as a bridge between understanding human-AI interaction and designing inclusive technologies around users' own linguistic priorities.

Abstract: 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: While recent advances in text-to-speech (TTS) synthesis have substantially improved speech quality and modeling flexibility, the development of lightweight and computationally efficient systems remains comparatively underexplored. We present LeanTTS, a bilingual complete synthesis stack that supports two complementary regimes: reference-prompt zero-shot cloning for unseen speakers and token-conditioned customization for recurring voices. LeanTTS combines a compact Mandarin polyphone front end, a structurally compressed flow-matching acoustic model, and a lightweight Vocos decoder with four-step MeanFlow generation. The resulting system contains 18.9M learned parameters, 86.12% fewer than the 136.2M-parameter ZipVoice-Vocos pipeline. On Seed-TTS-Eval, its zero-shot mode obtains 2.02 CER/62.3 SIM on Mandarin and 2.32 WER/48.3 SIM on English, exposing the quality-size trade-off under aggressive compression. With token-conditioned customized voices, the same system reaches 1.57 CER/77.2 SIM on Mandarin and 1.93 WER/76.6 SIM on English, with UTMOS scores of 3.03 and 4.33, respectively. It further achieves mean/P90 CPU RTFs of 0.049/0.061, establishing a sub-20M operating point for bilingual personalization and real-time CPU deployment. Audio demos are available at https://github.com/notlu/leantts.

Abstract: Unified streaming speech language models aim to support speech recognition and synthesis within a single model, enabling real-time human-computer interaction. However, jointly supporting these two directions of speech-text conversion while preserving task-specific performance remains challenging. We propose Uni-Stream, a unified streaming speech language model built on compact pretrained speech models. Uni-Stream processes interleaved text and speech in a streaming format and achieves performance comparable to task-specific models through two-stage training. We also develop a data annotation pipeline for streaming training and will publicly release the generated transcriptions and word-level timestamps. We further examine Uni-Stream's representations across tasks and modalities and find that speech and text residence in distinct latent feature space even when trained in a unified model architecture, while representations within each modality become more task-dependent in later layers.

Abstract: Zero-shot cross-lingual voice cloning often carries the reference speaker's accent into the target language. We propose closed-loop dual-reference guidance, which pairs the original reference with a reusable native anchor constructed by voice conversion (VC). The references initialize separate language-model (LM) states that guide a shared acoustic trajectory; each completed patch then updates both states, keeping their generated histories aligned with the emitted speech. Anchor construction runs once per speaker, and online synthesis uses a frozen TTS model without per-utterance VC. On 20 held-out French speakers, the fixed setting $λ^{\!*}=0.33$ reduces residual French LID mass by 73\% versus direct cloning, at an ECAPA similarity cost of only $0.042$; relative to anchor-only cloning, it recovers $0.155$. The method achieves the lowest mean WER ($0.009$) and highest mean GOP ($4.750$) among the held-out test arms. In blinded preference tests, native-English listeners favor it over direct cloning on pronunciation naturalness, and over both anchor-only cloning and a TTS$\to$VC cascade on speaker identity. Online RTF is approximately 26\% lower than the cascade's ($0.59$ vs.\ $0.80$), and the same setting reduces Spanish LID mass by 56\% without re-tuning.

Abstract: Speech separation (SS) has witnessed substantial progress in recent years with the advent of neural network-based methods, achieving notable improvements on signal-level evaluation metrics. Existing speech separation systems are mostly optimized with signal-level objectives, and often yield outputs that are acceptable in waveform metrics but suboptimal for intelligibility-sensitive downstream tasks. To address this gap, we propose SLM-SS, a token-based generative speech separation framework that models multi-talker mixtures with speech language models. We first discretize clean speech into multi-codebook token sequences with neural codecs and serialize multi-speaker targets via Serialized Output Training (SOT). An encoder-decoder autoregressive speech language model predicts the zero-order sequence to establish speaker ordering and coarse linguistic content, and a non-autoregressive module then refines higher-order residual codebooks for efficient reconstruction. We further introduce a per-talker non-autoregressive strategy that simplifies high-order modeling and improves the stability. On Libri2Mix, SLM-SS with per-talker NAR achieves $6.92\%$ WER, approaching codec-only reconstruction. The WavTokenizer variant eliminates NAR refinement and achieves the highest mean subjective quality score among evaluated separation systems. Its UTMOS and DNSMOS advantages over discriminative baselines persist on unseen noisy and Chinese mixtures without fine-tuning.

Abstract: The real-world usability of ASR models remains challenging, particularly for low-resource languages, where existing speech resources are often collected in clean studio environments. In this work, we introduce TeleASR-Amharic, a benchmark for evaluating Amharic ASR under various acoustic and telecommunication conditions. We analyze the major degradations observed in real telecommunications recordings and develop a benchmark that covers multiple channel conditions, environmental disturbances, and telecom-domain variations. Our results show that telecommunication degradation can substantially reduce recognition performance, making even high-performing ASR models impractical under severe conditions. For example, the WER of the best-performing model increases from 0.23 to 0.79 under severe telecommunication degradation, corresponding to a relative increase of 243% in recognition errors. We further investigate the effects of different noise and channel distortions and show that scaling training data and applying carefully designed telecommunication-aware augmentation substantially improve robustness. In particular, our best training configuration achieves a 23.6% relative WER reduction under extreme noise conditions. Together with the release of our telecommunications-domain data, these findings provide a reproducible benchmark for measuring progress in Amharic ASR while highlighting open challenges in developing robust speech technologies for real-world deployment.

Abstract: Large self-supervised speech models provide powerful representations for speaker verification, but full fine-tuning is expensive. Most conventional adapters insert small trainable modules to modify the output of each layer, while keeping the original residual connections unchanged. Building on our preliminary study of manifold-constrained Hyper-Connections in speaker encoders, we explore parameter-efficient adaptation of both residual paths and feature selection across layers. We first propose a Manifold-Constrained Residual Adapter (MCRA), which splits the hidden dimensions of each Transformer residual state into N groups and uses a small Sinkhorn-normalized matrix to exchange information among them. A learnable gate controls the balance between the original and mixed residual paths. The gate adds only one scalar parameter at each residual connection and initially favors the original path, allowing gradual adaptation. We further propose Frame-Adaptive Layer Aggregation (FALA), an attention-based aggregation method that uses a single learned query shared across all utterances to compute layer weights separately for each speech frame. Because the layer representations depend on the input, the resulting weights vary across both utterances and frames, unlike the fixed weights of a conventional layer-weighted sum. Compared with a frozen WavLM baseline using a fixed layer-weighted sum, our method achieves relative EER reductions of 22.4% on VoxCeleb1-O and 2.9% on CN-Celeb, with trainable adaptation parameters equivalent to only 0.0009% and 0.0016% of the WavLM-Large parameters, respectively.

Abstract: The previous version contained Unicode em dashes. Paste this ASCII-only version: Game theory identifies when a strategic action should be accepted, but not how long a live, language-based counterpart will continue to accept it. We entered the GLEE competition in its final week with one protected agent, Milton, and four instrument agents that ran deterministic, per-game randomized trials of policy changes; Milton received changes only through pre-specified gates or registered, rollback-guarded production rules. In negotiation, learned pricing improves adjusted percentile by 0.071 (n=21,168); in bargaining, a continuation model conditioned on the rejected offer did not improve on a recency-weighted incumbent and was retired. In persuasion, a preregistered five-dose experiment (7,391 assigned games, blinded until a fixed stop) shows that recommending every unit raises sales by 0.194 of the maximum (95% CI [0.147, 0.241]) where the receiver's slack makes obedience rational (52% of the Bayesian benchmark), lowers them where it does not (-0.067), and produces a larger early-to-late purchase decline than honest recommendations (-0.123). A rule fixed before unblinding now routes Milton's dose per cell. Milton stood 11th of 460 agents in bargaining at the fixed stop; we also report the close, where the final day's pool shift moved all five agents' bargaining ratings together while their play was unchanged, and three dispatch defects were found by checking delivered behavior against the design.

Abstract: Speech recognition in real time in far-field conversational settings still incurs significant errors. Identifying potential error spans during streaming ASR permits a recourse mechanism where the speaker can be asked to repeat the affected content under cleaner acoustic conditions. We propose \sysname, a new framework for online error localization, that identifies spans likely to benefit from this intervention. Our supervision uses clean audio to estimate how the recognizer's output might change under improved acoustic conditions. We probabilistically align noisy and clean decodes using a pair-HMM to obtain soft supervision. These soft targets train a counterfactual gain scorer by combining a suite of acoustic and language features to produce token-level scores, that are further grouped into spans for targeted recourse. An LLM uses the ASR transcript of the repetition to revise only the selected span. At a precision of $70\%$, \sysname achieves the highest recall by identifying up to 52\% more erroneous tokens compared to the strongest baseline. Targeted repetition yields relative WER reductions of up to 24\% under a budget that accounts for both interruption frequency and the lengths of re-uttered spans.

Abstract: Every multilingual speech pipeline begins with a language decision, an error there propagates silently through recognition, translation and routing. In India it spans 22 scheduled languages across four language families, heard over short, noisy, narrowband and code-mixed audio. Foundational speech models increasingly present language identification (LID) as an inherent latent, generalized capability. Autoregressive decoders in speech models emit a special language ID token, and modern LLMs with audio capabilities can name the language when prompted. We test this assumption on identical clips and find that whisper and similar speech models has tokens for only 15 of the 23 languages, on one second inputs Gemini 3.5 Flash reaches 0.52 macro-F1 and GPT-audio 0.14. We present Anuvani LID, a dedicated 115M parameter LID conformer model trained on 54,643 hours of Indian speech spanning 22 Indian languages and Indian accented English. Along with designed and controlled studies on 102 condition benchmark with speaker and corpus held out tests. Across 37 conditions, it reaches 0.914 mean macro-F1, 5.2 points above Gemini 3.5 Flash and ahead of 14 other systems. It is 24 points ahead of Gemini at one second, scores 0.897 on all 23 languages where no other system exceeds 0.713, and is calibrated to an ECE of 0.008. Reliable Indian LID still requires a robust, dedicated and calibrated model.

Abstract: Full-duplex speech systems must decide when to enter, hold, or relinquish the conversational floor, even as multiple speakers overlap or backchannel without taking it. Existing evaluations study dyadic turn-taking or end-to-end agent behavior, but lack a unified, controlled evaluation of genuine floor transitions against confusable non-events in multi-party settings. We introduce MultiPartyTurnBench (MPTB), a benchmark of 240 synthetically generated, channel-separated conversations spanning two to five speakers and six interaction types. MPTB evaluates two complementary tasks: End-of-Turn (EOT) detection distinguishes genuine turn completion from temporary holds, while Interruption (INT) detection distinguishes successful floor takeovers from non-floor-taking speech. Exact event-conditioned annotations enable causal evaluation using false-positive-constrained recall and signed latency. Across acoustic, semantic, and turn-projection baselines, we find a stark asymmetry: learned systems achieve 82.5-92.6% EOT recall at 4.3-11.8% false-positive rates, while INT detection suffers from missed takeovers or excessive false activations. This asymmetry extends to five full-duplex agents which achieve 86.1-95.8% balanced accuracy on turn entry, yet perform near chance (46.9-54.7%) on selective interruption handling. Gemini-3.8-Live yields to 97.7% of backchannels, whereas MiniCPM-o-4.5 yields to only 1.4% of genuine interruptions. These findings establish selective floor control as a distinct, unresolved challenge for reliable multi-party voice interaction.

Abstract: We study finite stochastic partial monitoring (PM) games with side information, a paradigm of sequential learning with partial feedback. On each round $t \in [T]$, both player and environment observe a context and player chooses an action and environment simultaneously chooses an outcome. The player neither observes the outcome nor the loss; instead, they only observe a feedback symbol which is partially informative about the outcome. We utilize the Estimation-to-Decisions (E2D) meta-algorithm of Foster et al.(2021) to design E2D.PMSide, a computationally efficient framework for contextual partial monitoring, which uses an online regression oracle to efficiently estimate the outcome distributions based on historical (context, action, feedback symbol) tuples. Since the exploration rule given by the original E2D algorithm is intractable to compute, we propose two computationally efficient exploration rules that achieve optimal regret orders in $T$: in the globally observable setting, our exploration rule is based on $ε$-greedy, while in the locally observable setting, our exploration rule uses inverse gap weighting (Foster & Rakhlin, 2020), which was previously only known to be effective in the bandit setting. We show that a variant of E2D.PMSide named OE2D.PMSide can be combined with offline regression oracles to give an efficient algorithm with optimal regret when the contexts are drawn independently from a fixed distribution. We evaluate our algorithm on a synthetic label-efficient prediction game and an LLM routing task, in which a learner decides from prompt, uncertainty, and internal state features whether to escalate a query from a weak model to a strong one. We show that our method has superior performance to prior approaches.

Abstract: Speech-to-speech translation should preserve not only what is said, but also who says it and how it is said. Existing end-to-end S2ST, speech MLLM, and cascade systems can preserve content or timbre, but often struggle to realize emotion, tone, rhythm, pauses, and pragmatic attitude in the target language. We propose PACT, a target-space pragmatic prompting framework for expressive and voice-preserving S2ST. PACT extracts source pragmatic cues, projects them into the target-language expression space, calibrates the projected representation to a realizable target subspace, and converts it into continuous prompts for a frozen speech generator. On bidirectional Spanish-English TOPI evaluation, PACT achieves the best overall performance among evaluated systems across content fidelity, voice preservation, and expression preservation metrics.

Abstract: End-to-end speech-to-text retrieval is becoming an important capability for voice agents, allowing spoken queries to retrieve relevant documents without intermediate transcription. Training these retrievers typically depends on large collections of spoken query-document pairs that are costly to acquire and difficult to extend across domains and languages. We introduce \ours, which replaces such task-specific supervision with speech-transcript pairs from standard ASR corpora. A lightweight adapter projects the output of a frozen speech encoder into the retrieval space of a frozen pretrained text embedding model. \ours is competitive with cascaded ASR-then-retrieve systems while reducing inference latency, and outperforms existing publicly available speech embedding models. It also transfers to text-to-speech and speech-to-speech retrieval without additional training. Controlled analyses validate the speech-transcript alignment paradigm, demonstrating robustness to reduced training data and consistent gains from larger corpora. Further ablations of loss functions and trainable modules establish practical guidelines for the framework. Model weights and codes are to be released.

Abstract: Full-duplex spoken language models (FD-SLMs) support simultaneous listening and speaking, but often lag behind half-duplex spoken language models (HD-SLMs) in general capabilities. We propose Dialogue-to-Duplex (D2), a framework that converts pretrained HD-SLMs into high-capability FD-SLMs at fine temporal granularity. D2 adapts the original audio encoder to streaming through decoder-free representation distillation and learns full-duplex interaction through supervised fine-tuning. We systematically study the effects of chunk size on full-duplex interaction, general capabilities, and inference computation. We apply D2 to two distinct HD-SLM architectures, Qwen3-Omni and LLaMA-Omni 2. At a 12.5Hz update rate, Qwen3-Omni-D2 achieves 99.2% turn-taking success with 521-ms latency, 99.5% interruption success, and a state-of-the-art VoiceBench score of 70.1.

Abstract: Whisper-style ASR models provide strong zero-shot transcription, but their residual errors are not evenly distributed across datasets, recording conditions, and metadata-based groups. We study when lightweight low-rank adaptation (LoRA) improves Whisper and whether observed gains require group-specific adapters. Across Common Voice, LibriSpeech, and VoxPopuli, we compare zero-shot Whisper, pooled LoRA, and group-specific LoRA under matched within-group splits. We find that adaptation is not uniformly beneficial: gains are concentrated in difficult Common Voice groups, are consistent on VoxPopuli, and largely disappear on low-error LibriSpeech. Crucially, pooled LoRA often matches, and sometimes clearly outperforms, group-specific LoRA. Gains over zero-shot should therefore be tested against matched pooled adaptation before being interpreted as evidence of personalization.

Abstract: Nonverbal vocalizations (NVVs), such as laughter and sighs, convey meaning alongside words. NVV-aware speech systems must recognize and generate these events while preserving lexical accuracy and perceptual quality. Yet differences in annotation, resources, and evaluation complicate comparisons, and aggregate scores can obscure specific weaknesses. We review ten system descriptions from the NVVSpeech Challenge 2026, covering understanding and generation in Chinese and English across 16 event categories. Our analysis examines supervision construction, task adaptation, and inference-time or preference-based feedback. Results reveal persistent recognition difficulties with breath and gasp and trade-offs between event recognition and lexical accuracy, as well as between generation intelligibility and acoustic quality, with human listening providing complementary evaluation evidence. These findings identify priorities for advancing NVV-aware speech understanding, controllable generation, and comparable evaluation.

Abstract: Speaker anonymization (SA) aims to conceal speaker identity while preserving linguistic content and emotion, but existing streaming systems often lose emotion information through discrete content bottlenecks or require emotion-specific modeling. We present LiveAnon, a voice-conversion-based streaming SA system that directly conditions an autoregressive codec language model on hidden features from a pretrained streaming ASR encoder, preserving emotion without emotion-specific supervision. To reduce codec-generation latency, we introduce capped RVQ decoding, which preserves dependencies among coarse codebooks while predicting the remaining codebooks in parallel. Under VPC24, LiveAnon improves intelligibility and emotion preservation over fixed-target Stream-Voice-Anon (SVA), a strong prior streaming SA system, matches emotion-prompted SVA in emotion preservation, and achieves the strongest semi-informed privacy among streaming systems. Under VPC26, it ranks second across all metrics among the evaluated anonymization systems, demonstrating a strong privacy-utility balance.

Abstract: Large language models can post-edit ASR by using linguistic context to compare N-best hypotheses and rewrite the transcript when needed, often improving WER without retraining the recognizer. However, prompt-only correction is brittle: it may over-edit when top-1 is correct and hallucinate when the reference is absent, and often needs very large models. We propose oracle-guided fine-tuning that turns N-best lists into supervision via three curated pools: corrections (reference in the list but not top-1) to teach selection, regularizers (top-1 correct) to discourage unnecessary rewrites, and hard negatives (reference absent) to suppress hallucination. Fine-tuning on this data provides more effective supervision than random subset fine-tuning for learning a hybrid select-or-rewrite policy, with the clearest gains from around 3B parameters on the proprietary dataset but less conclusive results on CommonVoice, a divergence suggesting that public read-speech text may overlap with LLM pretraining and mask genuine post-editing gains.

Abstract: Recent accent classification systems can achieve high accuracy in standard in-corpus evaluations. However, when speakers overlap between the training and test sets, such results may overestimate model robustness compared with real-world deployment scenarios. This study revisits the robustness of DistilHuBERT-based accent classification by using L2-ARCTIC and multiple external English accent datasets. The comparison covers file-level and speaker-disjoint splits, along with balanced external-data training and teacher-student pseudo-labeling. The large performance gap between in-corpus and external evaluations suggests that file-level evaluation substantially overestimates performance, while speaker-disjoint evaluation can provide a more realistic estimate of the generalization to unseen speakers. In this study, models trained on external data alone do not reliably generalize to L2-ARCTIC. Among the tested configurations, balanced external-data training offers the most promising supervised setting. The pseudo-labeling experiments further indicate that confidence-based filtering alone is insufficient under domain mismatch. Overall, this study suggests that robust accent classification evaluation should simultaneously include speaker-disjoint evaluation and cross-corpus testing.

Abstract: Speech anonymization is evaluated with automatic speaker verification attackers whose performance depends on the attribute distribution of both their training and evaluation data, which may bias privacy estimates for some speaker groups. To alleviate this issue, we introduce VoxAnon, a dataset built from VoxCeleb1 and 2 with training and evaluation splits balanced across gender, age, accent, and pitch. We extract and cross-check attributes using two complementary methods, balance trial lists per subgroup, and draw test/enrollment utterances from distinct video sources in order to limit linguistic and session bias. Comparing VoxAnon with unbalanced training and test data across four attackers and three anonymizers, we show that attribute imbalance changes which subgroups appear hardest to re-identify, in directions that depend on the attribute. In the balanced setting, residual disparities persist, mostly tied to accent and pitch, and depend on the anonymizer/attacker pair. Code and trial lists are publicly available.

Abstract: A conversational speech model has to decide, at every pause, whether to speak. Existing evaluations test boundary detection, policy adherence, or aggregate statistics, not this decision. We place real-time speech systems at 427 pauses and 482 turn ends from 176 unscripted conversations in Seamless Interaction, give each the conversation so far, and score what it generates against what the person did: whether and how soon it speaks at a pause; its latency, length, and wording at a turn end; and whether it can answer a question about the preceding 30 to 120 seconds. Across six systems, the hosted APIs (GPT Realtime, Gemini 3.1 Flash Live, Grok Voice) respond at almost every pause whichever endpointer they use; only their latency, about 1 to 4 s with silence-based endpointers and about 8 s with semantic turn detection, decides whether the words land inside the pause. The full-duplex models (PersonaPlex, Moshi) are the only systems that regularly stay silent and the only ones that interrupt within a human reaction time. Hosted systems reply longer than the person. Given the history as text, hosted systems recall 41 to 60% of conversation facts under a Claude Opus 5.5 judge and 19 to 26% under a strict word matcher; hearing it as audio, full-duplex models recall 17 to 22% and 12 to 15%. We release the labels, points, harness, and every output.

Abstract: Recent full-duplex spoken dialog models enable simultaneous listening and speaking, but fine-grained models still advance their backbone autoregressively at every interaction frame. We introduce DiffuPlex, a rolling masked diffusion framework that reduces this sequential computation by predicting multiple future user and assistant frames in a single backbone wake. DiffuPlex consumes only a confident prefix of each predicted future while interaction continues at the original frame rate. As user speech arrives, it checks the corresponding user predictions and, when the interaction diverges, preserves already played assistant content while revising only the unplayed future. We consider two inference policies over the same predictor: DiffuPlex-LISTEN consumes multiple future frames when they predict assistant silence, whereas DiffuPlex-SPEAK can also consume predicted assistant speech. Across full-duplex interaction and spoken-language evaluations, DiffuPlex substantially reduces sequential backbone computation while largely preserving interaction behavior and general capability. DiffuPlex-LISTEN and DiffuPlex-SPEAK achieve $1.46\times$ and $1.59\times$ deployment-path wall-clock speedups and $1.61\times$ and $1.80\times$ Core LM speedups, with all measured backbone invocations completing within the 80ms interaction interval. Human evaluation shows that LISTEN preserves speech naturalness and conversational quality, while SPEAK retains conversational quality with some degradation in speech naturalness.

Abstract: Speaker verification (SV) performance degrades under language mismatch due to the entanglement of speaker identity with language-specific acoustic cues. To address this problem, we leverage large-scale self-supervised pre-trained models (PTMs) to learn language-agnostic speaker representations. We utilize PTMs as robust front-end feature extractors, capitalizing on their rich acoustic and linguistic knowledge acquired from vast, diverse audio data. These generalized features are then used to train a downstream speaker embedding network, effectively disentangling speaker identity from language-specific characteristics. We validate our approach on the TidyVoice2026 benchmark, which benchmarks SV under language mismatch. Our proposed system (team T02) achieves equal error rates (EERs) of 2.21% on tv26_eval-A and 2.99% on tv26_eval-U.

Abstract: As human--AI interactions become more conversational, full-duplex speech language models capable of natural real-time dialogue are growing in importance. Beyond generating appropriate responses, these models must coordinate turn-taking, backchanneling, and floor management in real time. Reinforcement learning (RL) provides a way to refine these behaviors through direct feedback on interaction outcomes. However, existing RL methods either apply timing feedback to a token policy or optimize semantic content, leaving the joint improvement of timing and content unresolved. We introduce HiPLEX, an RL framework that factorizes a pretrained full-duplex text policy into a control policy that decides when to emit content and a conditional content policy that decides what to emit. The first factor selects among 'pad', 'epad', and 'con'. The second selects a token only when 'con' is chosen. This hierarchy describes conditional actions within each frame and uses the model's existing text head. We route timing advantages to the token-group factor through event-causal masks derived from generated speech episodes, and route an LLM-judge semantic advantage to the conditional content factor. Across three Moshi seeds on Full-Duplex-Bench v1, HiPLEX reduces takeover rates during natural user pauses and backchannel opportunities, and shortens post-interruption response latency relative to GRPO, while maintaining comparable judged interruption-response quality. On Moshi and PersonaPlex, HiPLEX better matches pooled human turn-timing and backchannel-rate marginals than GRPO.

Abstract: Modern text-to-speech (TTS) systems can clone a target speaker from a short reference clip or be fine-tuned on a target voice, yet their behaviour on morphologically rich, lower-resource languages such as Turkish remain under-characterised. We present a systematic benchmark of four contemporary systems (Chatterbox, CosyVoice, OmniVoice, and VoxCPM2) evaluated across fine-tuned and zero-shot configurations, contrasted with a conventional VITS baseline and anchored to natural gold speech. Each configuration is scored with eighteen complementary objective metrics spanning learned naturalness predictors (UTMOS v2, DNSMOS-Pro, SCOREQ, WhisQA, AudioBox-PQ, NatScore, SpeechLMScore), intelligibility and signal-quality estimators (SQUIM PESQ/SI-SDR/STOI, Brouhaha), speaker similarity, distributional fidelity (TTSDS) and low-level acoustic descriptors. We further analyse how quality varies with utterance length and quantify long-form temporal consistency through speaker-identity and naturalness drift over chunked utterances. We release our evaluation code to support reproducible TTS evaluation.

Abstract: Real-time conversational agents require speaker diarization that streams and runs on a CPU. Most systems apply an utterance-level speaker encoder to short, heavily overlapping chunks, which wastes computation and leaves the model optimized for the wrong task. We instead turn a state-of-the-art speaker recognition architecture into a causal frame-level encoder that reads the stream once and emits one embedding every 80~ms from a bounded two-second window of past audio, and pair it with online clustering that gates every update on the self-similarity of the stream. Trained only by distillation from an utterance-level teacher on simulated and out-of-domain mixtures, and evaluated with one fixed set of hyperparameters, the system is the most accurate streaming diarizer on low-overlap benchmarks at sub-second latency, degrades far less than cache-based systems as the number of speakers grows, and runs five times faster than real time on a single CPU thread.

Abstract: Model intelligence and fast response jointly shape the quality of interaction with speech language models, yet remain difficult to achieve together. Explicit chain-of-thought (CoT) improves reasoning and audio understanding, but generating intermediate reasoning tokens delays responses. Describing fine-grained acoustic cues further lengthens CoT and increases latency. Latent reasoning can reduce this overhead, yet existing methods often trail CoT and remain limited by single-path supervision and reasoning budgets that do not adapt to problem difficulty. We introduce AURAL, which models a distribution over multiple plausible reasoning continuations in latent space and jointly predicts chunks of future states to reduce sequential forward passes and reasoning latency. To provide initial supervision for latent reasoning, we construct AuralReason-683K: 683K bilingual speech utterances (about 1,000 hours) with concise CoT for emotion recognition, empathetic dialogue, and general reasoning. AURAL-RL then explores beyond these traces, rewarding concise reasoning that yields high-quality answers and adapting reasoning effort to each problem. Across two backbones, AURAL-RL achieves performance comparable to CoT-RL, with larger gains over the respective supervised checkpoints on most metrics. Analysis further shows that harder questions elicit more latent reasoning steps. On Qwen2.5-Omni, it reduces time to the first answer token by 11.8x, from 1.22 to 0.10 s, versus 0.05 s for direct answering.

Abstract: Speech transcripts used as long-term memory must preserve both words and stable speaker identities. Existing meeting-transcription metrics either ignore speakers or remap anonymous speakers independently in each recording, so they cannot measure whether the same person retains one identity across meetings. We evaluate persistent speaker attribution with Speaker Identified cpWER (SI-cpWER), which scores a corpus under one global speaker-ID assignment. The benchmark covers five commercial diarize-then-identify cascades, two open academic baselines, and ThyVoice on the full 129-meeting CHiME-8 NOTSOFAR evaluation set in clean and noiseaugmented form, plus CHiME-6. ThyVoice is our end-to-end reference system; it repairs overlap and gates the evidence used to create and update voiceprints. Requiring persistent identity changes the commercial ranking: ThyVoice records lower SI-cpWER than every evaluated commercial cascade in all three conditions and the lowest mean in the full panel, 47.13 versus 54.75 for the next system. Complementary lexical, diarization, per-recording attribution, and speaker-clustering diagnostics characterize upstream error surfaces in the final attributed record. These results show why persistent attribution must be evaluated directly in systems that reuse conversations across time.

Abstract: Real-time voice conversion (VC) systems commonly rely on pretrained speaker embeddings from automatic speaker verification (ASV) models. While effective for speaker discrimination, these embeddings are trained to remain stable across phonetic and prosodic variations within-speaker, which may conflict with frame-level acoustic generation in streaming constraints. To address this issue, we propose VOSSA (Voiceprint Optimization for Streaming Speech Architectures), a speaker representation framework that extracts speaker information from intermediate content encoder layers and aggregates using attentive statistics pooling. The embedding is trained jointly with VC objectives, removing the need for a separate speaker encoder. Across six datasets, VOSSA improves F0 dynamics and vowel-discriminative acoustic cues while maintaining comparable NISQA-MOS, WER, and speaker similarity. Perceptual tests further indicate improvements in naturalness, speaker similarity, intelligibility, and vibrancy.

Abstract: Voice agents in production must handle several requests, background speech, and customers who lose patience. We introduce VAmoS Energy, a benchmark that combines these challenges in 100 calls about utility billing and payment assistance. Each caller makes two to four requests. The agent has sixteen tools backed by a stateful Stripe billing twin and the Apache Fineract loan engine, with account access blocked until caller verification succeeds. The tasks use public household electricity data and a policy based on Pennsylvania's residential billing rules. An LLM-as-a-verifier checks the agent's actions and spoken figures against explicit requirements. On a calibration run, it agrees with a code verifier on 99.1% of checks. Across fourteen voice stacks and three repeats per task, completion ranges from 17.3% to 44.7%. Grok Voice leads, and Gemini 3.8 Live and GPT-Live follow at about the same cost per call. Background television reduces pooled completion from 38.7% to 8.6%. The simulated caller often accepts an incorrect result because it hears the agent's words but cannot inspect its actions. These findings show why voice agents need evaluation across the whole call, including what they say, what they change, and how they handle competing speech.

Abstract: Speculative decoding accelerates autoregressive automatic speech recognition (ASR), but the acceptance length of alignment-aware drafters can saturate as the draft horizon increases. We propose a progress-aware speculative drafter that recurrently propagates an acoustic progress state across draft steps and feeds it back into audio cross-attention to guide token generation. We jointly train the drafter and progress predictor over variable draft horizons. On five ASR test sets, our method achieves lossless, macro-averaged end-to-end speedups of 1.657x and 1.227x over target-only autoregressive decoding with Qwen3-ASR-0.6B and Qwen3-ASR-1.7B, respectively. Relative to AnchorDraft, our method improves the macro-averaged speedup by 34.3% and 9.0%, respectively. Horizon sweeps show continued growth in acceptance length beyond the baselines' saturation. Code is available at https://github.com/yuanyuanjia71-spec/ProgDraft.

Abstract: Audio emotion recognition (AER) typically assigns a single label to an entire recording, leaving the temporal scope of that label ambiguous when multiple speakers and affective events are present. We address this limitation by reformulating AER as a Temporal Affective Grounding (TAG) task that associates emotions with temporally bounded speech spans and vocal tone descriptions. To support this formulation, we curate temporally annotated versions of existing emotion recognition datasets and construct recordings containing two to four affective speech spans, including overlapping speech. Training in this longer format with a standard language-modeling objective can degrade both emotion recognition and temporal grounding performance, while tone descriptions can provide shortcuts for emotion prediction. To address these challenges, we introduce Masked Temporal Affective Grounding (M-TAG), a supervised training objective that combines full-sequence language modeling with emotion and timestamp cross-entropy losses under attention masking. The masking varies the context visible to emotion-label tokens to reduce reliance on shortcuts and improve generalization, while the timestamp loss incorporates a distance-aware weight to penalize larger temporal errors. We evaluate EMO-TAG, a model fine-tuned using our dataset and objective, on emotion recognition and affective temporal-grounding against three AER and audio-language baselines: Flamingo-Next, Audio-Reasoner, and AffectGPT. Our results show that existing models achieve limited affective temporal-grounding despite competitive emotion recognition performance.

Abstract: Spoken conversational systems must recover information from prior interactions (i.e., memory), yet relevant information in speech extends beyond what was said to who said it, how it was spoken, and what was audible, information that exists only in the audio signal and cannot be recovered from a transcript. Beyond what to remember, memory also demands diverse operations: retrieving a single fact, integrating evidence across turns, tracking an evolving state. Real interactions further unfold across sessions, meaning information accumulates across distinct episodes rather than a single continuous recording. Existing benchmarks fall short on all three dimensions: they focus primarily on lexical content, adopt limited and ad hoc memory operations, and treat memory as a single-session problem. We argue that principled memory evaluation requires jointly characterizing the acoustic evidence to be retained and the operations applied to it, and introduce a taxonomy along these two axes. Building on this taxonomy, we present VoxMem: 3,196 evaluation instances over 34,743 spoken sessions (177 hours) crossing four acoustic evidence types (speech semantics, speaker identity, paralinguistic cues, environmental sound) with four memory operations (information extraction, multi-session reasoning, temporal tracking, and answer refusal), grounded in multi-session histories and stratified across context budgets from 8K to 64K tokens. Evaluating 15 LALMs, no model exceeds 40% at 32K. Models retain what was said far better than who said it, how, or what was audible, a gap that widens for complex operations, grows with history length, and manifests as qualitatively distinct failure modes across evidence types. VoxMem aims to provide a foundation to measure and drive progress on the full scope of spoken conversational memory.

Abstract: Many voice interaction applications require exact control over both the content and delivery of responses, typically using pre-recorded lines. Recent full-duplex models respond with low latency but cannot guarantee exact content or reproduce a specific recorded performance, while cascaded systems can be constrained to predefined responses at the cost of additional latency. We propose RePlay, a spoken dialogue system adapted from PersonaPlex that handles multi-turn conversations by retrieving and playing pre-recorded lines. Using probing, we identify the layer and frame at which the upcoming response becomes recoverable, and use this hidden state as the retrieval query. RePlay retains only the layers up to that point and replaces text and speech generation with lightweight turn-taking and retrieval heads. In simulated multi-turn interviews, RePlay reaches a median latency of 383 ms, 3 to 7 times lower than ASR-LLM cascades of comparable dialogue quality, at the cost of lower exact-line accuracy. In a user study, participants preferred RePlay in 63% of ratings versus 12% for a fast cascade with a small LLM (p = 0.008), and showed a non-significant preference (46% vs. 21%) over a slower cascade with a stronger LLM.

Abstract: Voice agents must complete users' tasks despite noise, reverberation, and competing speech. Evaluating agents' robustness therefore requires following how acoustic conditions affect the conversation and the actions taken on the user's behalf. This overview examines what existing benchmarks reveal about agents' ability to complete tasks under acoustic stress and where further task-based evaluation is required. We then introduce TRACE, a practical workflow for designing, running, and interpreting evaluations of acoustic robustness in task-oriented human-agent interactions: the same agent attempts a specified task with an original recording and an acoustically stressed copy, and the resulting conversations are scored for task completion, wrong actions, recovery, and user effort. Finally, we explain how results from these evaluations can guide changes to an agent to prevent wrong actions and improve recovery.

Abstract: This paper presents BanglaTurn, a corpus for end-of-turn detection in Bangla conversational speech, and a model trained on it. The corpus holds 35,374 samples of 3 to 15 s of podcast speech, labelled for turn state by combining speaker diarization with an LLM pass, with every label then checked by a human annotator. The model pairs a Whisper encoder with task-specific classification heads. On a class-balanced test set drawn from a held-out podcast, it reaches 84.33% accuracy (95% CI 80.3 to 88.1) against 69.28% for the Smart-Turn v3 baseline, and lowers the false negative rate from 51.57% to 7.55% at the cost of a higher false positive rate. We report what encoder layer fine-tuning, multi-scale pooling and INT8 quantization each contribute, and latency stays within 165 to 191 ms end to end on CPU.

Abstract: While Chain-of-Thought (CoT) reasoning has improved the capability of language models, directly applying it to Spoken Language Models (SLMs) may introduce long silent intervals under the serial "think-then-speak" paradigm, disrupting real-time spoken interaction. To address this issue, we propose an asynchronous think-aloud framework for reasoning-based SLMs within the Thinker-Talker architecture. The framework maintains a primary reasoning stream for logical deduction and a lightweight think-aloud stream that generates short, task-grounded progress utterances conditioned on the user input and the evolving reasoning state. A dynamic balance strategy coordinates the two streams at runtime, triggering additional think-aloud speech to avoid silent gaps and canceling pending utterances when the final response becomes ready. Experiments on spoken reasoning and question-answering benchmarks show that our approach substantially reduces user-audible silence during reasoning while maintaining answer accuracy comparable to that of a serial "think-then-speak" baseline, demonstrating the potential of asynchronous think-aloud for responsive interaction in SLMs.

Abstract: Detecting emotions is necessary for building systems that can accurately and adaptively interact with humans. Speech Emotion Recognition (SER) has become an important research focus to develop intelligent spoken interfaces. However, most studies predict emotions at the utterance level, ignoring the conversational context, along with the emotional flow and speaker interactions it carries. In this paper, we introduce ACERT (Averaged Contextual Emotion Representation through Time), a module that integrates a flexible-length window of conversational context to better capture emotional evolution in spoken interactions. To evaluate the robustness of this method, we conducted experiments on datasets spanning diverse emotionally expressive styles and contexts. ACERT outperforms current state-of-the-art (SOTA) approaches on IEMOCAP, establishes the first context-aware benchmark on SAFE, and obtains strong results on MELD for unweighted, class-balanced metrics. Ablation studies show that ACERT's gains come from emotional and conversational continuity, rather than from speaker identity or acoustic conditions.

Abstract: Instruction-controlled speech synthesis and paralinguistic understanding are often trained independently, leaving reciprocal feedback between the two tasks underexplored. We introduce CycleSpeech, a framework that connects generation and understanding through a shared, structured voice profile that serves as a common target for supervision and reciprocal feedback. The forward cycle assesses whether synthesized speech expresses the intended attributes by comparing recovered and target profiles. The backward cycle evaluates whether profiles inferred from real speech can guide reconstruction of the source speaking style. To support both directions, we construct a bilingual dataset of 20,046 examples pairing instructions, target speech, speaker references, and structured profiles. Building on joint supervised fine-tuning, CycleGRPO alternates policy updates using reciprocal rewards grounded in profile consistency and speaking-style reconstruction. Fixed target profiles anchor feedback from the evolving counterpart. This procedure requires neither human preference annotations nor an additional preference-trained reward model. Evaluations on Chinese and English benchmarks show improved instruction adherence and profile recovery while maintaining competitive synthesis quality. Compared with Step-Audio-2-mini, CycleSpeech improves instruction-match accuracy by 4.50 and 10.06 percentage points in Chinese and English, respectively. Controlled ablations further support the contribution of cycle feedback to generation control. These results support structured voice profiles as an interface for reciprocal training between speech generation and paralinguistic understanding. An online demo is available at https://cyclespeech.github.io.

Abstract: Evaluation benchmarks for Indian language automatic speech recognition (ASR) suffer from two systematic biases: optimistic scores from clean, controlled audio conditions, and pessimistic scores from overly rigid transcription standards that penalize valid linguistic variations. We introduce Vimarsha, a 100-hour benchmark spanning all 22 scheduled Indian languages, designed to address both distortions. Vimarsha combines demographically diverse on-field recordings with carefully mined in-the-wild audio selected for acoustic difficulty, alongside a lattice of variations framework that encodes multiple valid transcriptions per utterance. Evaluations of 10 state-of-the-art ASR models reveal substantial shifts in model rankings under realistic conditions, geographic and demographic performance disparities, and systematic failure modes across speaking rates and acoustic environments.

Abstract: Large Audio Language Models (LALMs) can follow diverse instructions to synthesize speech in specified styles. However, complex instructions that require simultaneous control over pitch dynamics, speaking rate, and emotional tone often exceed what a single-pass generation can faithfully realize. While recent reasoning models have shown that intermediate "thinking" tokens improve output quality, this paradigm has been confined to the text modality. In this work, we extend reasoning to the audio token space by training a LALM with reinforcement learning to reason over its own speech output. The model first generates a draft speech as a form of audio-token reasoning, critiques its own generation by reflecting on the acoustic realization in text, and then produces a refined version conditioned on both the first-pass speech and the critique, all within a single model. After RL training, the refined two-hop outputs achieve a relative improvement of 7.15\% on the InstructTTSEval benchmark, demonstrating the model's reflective ability.

Abstract: Recent advances in Audio LLMs have achieved human-level speech recognition, yet existing systems struggle to capture paralinguistic aspects such as speaker traits, expressive variations, and environmental acoustic conditions. To address this, we design a framework of 22 paralinguistic characteristics and create a dataset of over 1.2M Audio-QA pairs. We develop ParA-LLM, trained with a two-stage curriculum: first on single-attribute questions to build foundational knowledge, then on multi-attribute questions for joint reasoning over speaker and acoustic characteristics. We also release ParA-Bench, a benchmark of 6,000 multiple-choice questions across speaker-speech, acoustic, and mixed categories, where frontier models like GPT-4o-Audio achieve only 36% accuracy. ParA-LLM surpasses state-of-the-art Audio LLMs like GPT-4o-Audio by 7.5% on ParA-Bench, with additional gains of 1.13% on MMAU-Pro Speech and 7.49% on MMAR Speech.

Abstract: FarmerChat is Digital Green's AI-powered agricultural advisory assistant for smallholder farmers, who access it in their own language through text, voice, or photographs. Voice is a critical channel for this population, yet field-recorded speech is challenging for general-purpose automatic speech recognition (ASR) because recordings frequently contain machinery noise, background media, competing speakers, and domain-specific agricultural vocabulary. These conditions disproportionately affect crop, pest, chemical, and quantity terms that carry the meaning of a farmer's query. We present a modular, model-agnostic pipeline for improving ASR quality in FarmerChat without fine-tuning or replacing the underlying ASR model. The pipeline combines gated audio enhancement, speaker diarization and target-speaker selection, ASR, domain-aware correction using a weighted agricultural lexicon, and a quality gate for detecting unreliable transcripts. Only the diarization stage is fine-tuned; all other stages use off-the-shelf models behind common interfaces. We evaluate the pipeline on human-annotated FarmerChat recordings in Hindi, Telugu, and Odia using word error rate (WER) and a domain-weighted error rate that gives greater importance to agricultural terminology. The largest improvements occur on multi-speaker recordings, where target-speaker selection prevents competing speech from entering the transcript. Across the full corpus, the pipeline reduces WER by 16-23% relative on three cloud ASR models and by 5% on an on-device model. On multi-speaker recordings, the reductions are 32-42% for the cloud models and 16% for the on-device model. All reported reductions are statistically significant. These results show that targeted preprocessing, speaker selection, and domain-aware post-processing can substantially improve agricultural speech transcription while preserving the underlying ASR model.

Abstract: Pure speech language models often lag behind text and speech-text language models in generating coherent content, but this gap is difficult to quantify because speech and text systems are typically evaluated with different metrics and trained on different data. We study the speech-text modality gap in a family of spoken language models, based on flow matching for continuous acoustic feature generation. We construct a unified generation-based evaluation suite that compares speech-only, text-only, and speech-text language models trained on matched data distributions and evaluated in matched generation settings. We evaluate generated continuations along multiple dimensions: semantic coherence, measured by transcribing generated speech and scoring it with a reference language model; local phonetic structure, measured by phone n-gram distributional statistics; speaker consistency and acoustic quality; and emotion-based distributional metrics. Across datasets, we find that joint speech-text modeling substantially improves semantic coherence. However, the improvement is not uniform across metrics: phone-level metrics change only modestly, speaker similarity and predicted quality are lower for speech-text continuations, while emotion-based distributional metrics improve. Compared with larger-scale speech-only models, our speech-text model closes much of the scaling gap in transcript-based semantic coherence, suggesting that text provides an efficient semantic training signal for spoken language modeling.

Abstract: The ISCSLP 2026 CoT-TTS Challenge requires TTS systems to generate Chain-of-Thought (CoT) reasoning from dialogue history before synthesizing contextually appropriate speech. While the official baseline establishes a unified architecture, it remains constrained by limited contextual comprehension, weak instruction fidelity, and suboptimal audio quality. We present a systematic optimization pipeline to address these limitations. First, we develop a data process framework that cleans raw data via FullSubNet denoising, Qwen3-ASR re-transcription, and Qwen3.5-35B-A3B-based history-CoT consistency analysis, while distilling 545K high-fidelity instruction samples using Qwen3-TTS and Seed-VC under strict quality filtration. Second, we propose a Context-Aware Direct Preference Optimization (CA-DPO) method. By employing a cascaded filtering strategy, ASR prescreening, LLM tournament ranking, and speaker similarity verification, we obtain high-confidence preference pairs that significantly enhance holistic ``Context$\rightarrow$CoT$\rightarrow$Speech'' consistency during DPO training. Third, we establish an evaluation method featuring a 500-sample test set and an LLM-as-Judge framework to independently assess reasoning and execution fidelity. Experiments demonstrate that our system significantly outperforms the baseline across all objective and subjective metrics, validating our data governance and alignment strategies.

Abstract: We report a multi-month engineering program to build Sophea, a production bilingual Greek-English automatic speech recognition system. We evaluate the system against nine production gates covering Greek and English word error rate, language identification, and hallucinations on non-speech audio. Across twenty-three training iterations and two model architectures, no training-data composition passed all nine gates simultaneously. Meeting the Greek noisy-environment target required about 1,500 steps of dense domain exposure, while preserving English language identification tolerated only about 250 steps, or about 1,250 with a rebalanced mix that reduced Greek accuracy. We describe a six-stage data pipeline in which calibrating an audio-quality filter against in-domain anchors reduced the discarded share of scored Greek audio from 98.7 percent to 10.6 percent. A pre-registered ablation isolated a hallucination defect to one training-data package. A three-model ROVER ensemble increased gate coverage from 4-7 of 9 for individual models to 9 of 9 and reduced overlapping-speech WER from 53.35 percent to 37.87 percent, a 29 percent relative improvement. A separate learned per-clip arbiter over two models is listed as sophea/asr-k1 (preview) on the public Open ASR Leaderboard, with 4.26 percent average WER across eight public English test sets, and reaches 25.88 percent WER on live Greek noisy-environment traffic. We also document five cases in which a measurement tool produced a plausible but incorrect result and seven substantial approaches that were evaluated but not shipped. No model weights or training data are released; we report methodology and quantitative results only.

Abstract: In conversational AI, detecting when a speaker has finished talking is crucial for natural turn taking. While recent work incorporates semantics, the relative contribution of different modalities remains unclear. We present a controlled ablation of acoustic, prosodic, and semantic signals for streaming end of turn detection using a lightweight trimodal classifier. Under identical training conditions, the acoustic prosodic combination achieves the best balance of accuracy and latency, achieving utterance F1 of 0.93 with 7.8% false alarms at 400ms median latency. Adding text increases premature detections without improving performance. Feature space analysis confirms that prosodic features have the strongest class separability, while text representations overlap substantially. These findings suggest that turn-taking is primarily conveyed through intonation and silence patterns rather than semantic completeness, enabling faster and more reliable systems without expensive text inference.

Abstract: Turn-taking is a fundamental mechanism that governs when interlocutors speak and listen. Although Spoken Dialogue Systems (SDS) exploit a range of linguistic, acoustic, and non-verbal cues, they produce ill-timed responses in unscripted interaction. A central challenge is anticipating Transition Relevance Places (TRPs), or opportunities, not obligations, for a listener to take the floor. Human listeners do not wait for turn endings; as an utterance unfolds, they use expectations about its developing meaning to anticipate TRPs and decide whether to take the floor. We examine whether these evolving expectations can be modeled through semantic uncertainty -- an LLM-derived measure of how strongly a turn so far constrains what may plausibly come next. To do so, we sample possible continuations of an ongoing turn and use changes in semantic dispersion to identify TRPs within turns. We evaluate this account on a dataset with TRP labels derived from real-time listener responses, rather than retrospective annotation. Our approach substantially outperforms prompt-based and fine-tuned text-only baselines, providing empirical support for the view that evolving semantic constraints inform perceived turn-taking opportunities in unscripted interaction.

Abstract: Streaming multi-speaker ASR is a challenging task that must balance accuracy, latency, and efficiency while handling overlapping speech and maintaining coherent long-context modeling over extended conversations in an online fashion. We present a unified framework that categorizes streaming multi-speaker ASR into four architectural strategies based on how diarization and ASR are integrated. Using a shared pair of open-source streaming ASR and diarization models as a common foundation, we derive four multi-speaker ASR systems that differ in whether they employ multiple model instances, fine-tuning, or both. We evaluate these systems across multi-speaker accuracy, single-speaker accuracy degradation, memory footprint, and training complexity. Through this systematic architectural analysis, we clarify the design space for streaming multi-speaker ASR and provide practical guidance for selecting the most suitable approach under diverse deployment constraints.

Abstract: Human speech includes nonverbal vocalizations (NVVs), such as laughter, sighs, breaths, and coughs, which convey affective and interactional information. Existing approaches typically represent NVVs as transcript-level tags, providing limited supervision for their waveform-time boundaries. We present NVV-Locator for fine-grained NVV temporal grounding. We first unify 26 NVV categories across public resources and construct large-scale timestamp-supervised training data through dual-LLM verification, transcript-guided forced alignment, and energy-based boundary refinement. We further introduce NVV-TimeBench, an expert-refined benchmark with 667 utterances and 1,094 events. NVV-Locator uses a non-autoregressive slot-filling architecture to jointly predict lexical timestamps, NVV categories, and event boundaries. On NVV-TimeBench, it achieves 71.0% Micro F1, 70.2% Macro F1, 80.4% Macro mIoU, and 59.6 ms Macro mMAE, outperforming the evaluated large audio model counterparts. Evaluation on an external corpus further demonstrates the cross-corpus generalization of NVV-Locator.