arXiv subjects

Rini Sharon

Publications and source records attributed to Rini Sharon.

6 recordsLinked to original sources

arXiv2026

How Much Prosody Does Your ASR Encoder Already Know

End-of-turn (EoT) detection decides in real-time whether a speaker has finished. Traditional detectors rely on hand-built prosodic features (pitch, speaking rate, pauses), while modern detectors run on a neural ASR encoder that has processed the same audio. Because the encoder reads the signal those features are computed from, it may already hold much of the information they were meant to supply. We test this three ways, reading a streaming Zipformer encoder at four depths from raw input to the final layer. First, we measure how much prosody helps over the encoder alone. Second, we test whether prosody can be recovered from the encoder. Third, we ask whether any single feature still helps given the encoder. The three tests agree; prosody adds a large gain on raw audio, but that gain collapses once the signal passes through even the first encoder layer. The same pattern holds for two open-source encoders, wav2vec2 and Whisper, both of which are offline.

eess.AS↗

arXiv2026

The Importance of Speech Genre for End-of-Turn Detector Training

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.

eess.AS↗

arXiv2026

Towards Multilingual End-of-Turn Detection using Frozen Speech Encoders

A spoken dialogue system must judge when the user has yielded the floor. This end-of-turn (EoT) detector is normally trained and tuned afresh for each language, which is costly to scale to new languages. We instead build one on a frozen multilingual encoder, training only a small fixed-size head. Using five CALLHOME languages, we investigate three questions. First, how well a head trained on one language transfers zero-shot to the others. Second, how much target-language adaptation is required to recover full task performance. Third, whether a single fixed decision threshold holds across languages without per-language tuning. We find that the head transfers well, even between tonal and pitch-accent languages, and that a few hours of audio recover nearly all the loss, matching in-language training. A frozen multilingual encoder with one lightweight head is a practical, language-agnostic baseline for EoT detection.

eess.AS↗

arXiv2026

Less can be More: What Aspects of Speech Drive End-of-Turn Detection

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.

eess.AS↗

arXiv2022

Multilingual and Multimodal Abuse Detection

The presence of abusive content on social media platforms is undesirable as it severely impedes healthy and safe social media interactions. While automatic abuse detection has been widely explored in textual domain, audio abuse detection still remains unexplored. In this paper, we attempt abuse detection in conversational audio from a multimodal perspective in a multilingual social media setting. Our key hypothesis is that along with the modelling of audio, incorporating discriminative information from other modalities can be highly beneficial for this task. Our proposed method, MADA, explicitly focuses on two modalities other than the audio itself, namely, the underlying emotions expressed in the abusive audio and the semantic information encapsulated in the corresponding textual form. Observations prove that MADA demonstrates gains over audio-only approaches on the ADIMA dataset. We test the proposed approach on 10 different languages and observe consistent gains in the range 0.6%-5.2% by leveraging multiple modalities. We also perform extensive ablation experiments for studying the contributions of every modality and observe the best results while leveraging all the modalities together. Additionally, we perform experiments to empirically confirm that there is a strong correlation between underlying emotions and abusive behaviour.

eess.AS↗

arXiv2022

ADIMA: Abuse Detection In Multilingual Audio

Abusive content detection in spoken text can be addressed by performing Automatic Speech Recognition (ASR) and leveraging advancements in natural language processing. However, ASR models introduce latency and often perform sub-optimally for profane words as they are underrepresented in training corpora and not spoken clearly or completely. Exploration of this problem entirely in the audio domain has largely been limited by the lack of audio datasets. Building on these challenges, we propose ADIMA, a novel, linguistically diverse, ethically sourced, expert annotated and well-balanced multilingual profanity detection audio dataset comprising of 11,775 audio samples in 10 Indic languages spanning 65 hours and spoken by 6,446 unique users. Through quantitative experiments across monolingual and cross-lingual zero-shot settings, we take the first step in democratizing audio based content moderation in Indic languages and set forth our dataset to pave future work.

cs.SD↗