Neville Ryant

Seth Kulick, Neville Ryant, David J. Irwin et al.

This paper addresses the problem of improving POS tagging of transcripts of speech from clinical populations. In contrast to prior work on parsing and POS tagging of transcribed speech, we do not make use of an in domain treebank for training. Instead, we train on an out of domain treebank of newswire using data augmentation techniques to make these structures resemble natural, spontaneous speech. We trained a parser with and without the augmented data and tested its performance using manually validated POS tags in clinical speech produced by patients with various types of neurodegenerative conditions.

Seth Kulick, Neville Ryant, Beatrice Santorini et al.

We describe the construction and evaluation of a part-of-speech tagger for Yiddish. This is the first step in a larger project of automatically assigning part-of-speech tags and syntactic structure to Yiddish text for purposes of linguistic research. We combine two resources for the current work - an 80K-word subset of the Penn Parsed Corpus of Historical Yiddish (PPCHY) and 650 million words of OCR'd Yiddish text from the Yiddish Book Center (YBC). Yiddish orthography in the YBC corpus has many spelling inconsistencies, and we present some evidence that even simple non-contextualized embeddings trained on YBC are able to capture the relationships among spelling variants without the need to first "standardize" the corpus. We also use YBC for continued pretraining of contexualized embeddings, which are then integrated into a tagger model trained and evaluated on the PPCHY. We evaluate the tagger performance on a 10-fold cross-validation split, showing that the use of the YBC text for the contextualized embeddings improves tagger performance. We conclude by discussing some next steps, including the need for additional annotated training and test data.

Jonathan Wright, Mark Liberman, Neville Ryant et al.

A random sample of nearly 10 hours of speech from PennSound, the world's largest online collection of poetry readings and discussions, was used as a benchmark to evaluate several commercial and open-source speech-to-text systems. PennSound's wide variation in recording conditions and speech styles makes it a good representative for many other untranscribed audio collections. Reference transcripts were created by trained annotators, and system transcripts were produced from AWS, Azure, Google, IBM, NeMo, Rev.ai, Whisper, and Whisper.cpp. Based on word error rate, Rev.ai was the top performer, and Whisper was the top open source performer (as long as hallucinations were avoided). AWS had the best diarization error rates among three systems. However, WER and DER differences were slim, and various tradeoffs may motivate choosing different systems for different end users. We also examine the issue of hallucinations in Whisper. Users of Whisper should be cautioned to be aware of runtime options, and whether the speed vs accuracy trade off is acceptable.

Shinji Watanabe, Michael Mandel, Jon Barker et al.

Following the success of the 1st, 2nd, 3rd, 4th and 5th CHiME challenges we organize the 6th CHiME Speech Separation and Recognition Challenge (CHiME-6). The new challenge revisits the previous CHiME-5 challenge and further considers the problem of distant multi-microphone conversational speech diarization and recognition in everyday home environments. Speech material is the same as the previous CHiME-5 recordings except for accurate array synchronization. The material was elicited using a dinner party scenario with efforts taken to capture data that is representative of natural conversational speech. This paper provides a baseline description of the CHiME-6 challenge for both segmented multispeaker speech recognition (Track 1) and unsegmented multispeaker speech recognition (Track 2). Of note, Track 2 is the first challenge activity in the community to tackle an unsegmented multispeaker speech recognition scenario with a complete set of reproducible open source baselines providing speech enhancement, speaker diarization, and speech recognition modules.

Ilias Triantafyllopoulos, Young-Min Cho, Ren Tao et al.

Activation-based steering provides control of LLM behavior at inference time, but the dominant paradigm reduces each concept to a single direction whose geometry is left largely unexamined. Rather than selecting a single steering direction, we use conceptors: soft projection matrices estimated from activations pooled across both poles of a bipolar concept, which preserve the concept's full multidimensional subspace. A geometric analysis shows the bipolar subspace strictly subsumes the single-vector baseline. We further show that the conceptor quota provides a parameter-free layer-selection diagnostic, predicting concept separability with Pearson correlations up to r=0.96 across three instruction-tuned models and three semantic dimensions. Beyond selection, conceptors admit a closed-form Boolean algebra (AND, OR, NOT): we evaluate conceptor compositionality on thematically related sub-concepts. Across a systematic five-axis design-space evaluation, conceptors match or outperform additive baselines at layers where concept subspaces are multi-dimensional while producing substantially fewer degenerate outputs. Conceptor steering is a geometrically principled, compositional, and practically safer alternative to single-direction steering from a limited number of contrastive pairs.

Rajath Rao, Adithya Ganesan, Oscar Kjell et al.

Current speech encoding pipelines often rely on an additional text-based LM to get robust representations of human communication, even though SotA speech-to-text models often have a LM within. This work proposes an approach to improve the LM within an audio model such that the subsequent text-LM is unnecessary. We introduce WhiSPA (Whisper with Semantic and Psychological Alignment), which leverages a novel audio training objective: contrastive loss with a language model embedding as a teacher. Using over 500k speech segments from mental health audio interviews, we evaluate the utility of aligning Whisper's latent space with semantic representations from a text autoencoder (SBERT) and lexically derived embeddings of basic psychological dimensions: emotion and personality. Over self-supervised affective tasks and downstream psychological tasks, WhiSPA surpasses current speech encoders, achieving an average error reduction of 73.4% and 83.8%, respectively. WhiSPA demonstrates that it is not always necessary to run a subsequent text LM on speech-to-text output in order to get a rich psychological representation of human communication.

Seth Kulick, Neville Ryant, Beatrice Santorini

We present the first parsing results on the Penn-Helsinki Parsed Corpus of Early Modern English (PPCEME), a 1.9 million word treebank that is an important resource for research in syntactic change. We describe key features of PPCEME that make it challenging for parsing, including a larger and more varied set of function tags than in the Penn Treebank. We present results for this corpus using a modified version of the Berkeley Neural Parser and the approach to function tag recovery of Gabbard et al (2006). Despite its simplicity, this approach works surprisingly well, suggesting it is possible to recover the original structure with sufficient accuracy to support linguistic applications (e.g., searching for syntactic structures of interest). However, for a subset of function tags (e.g., the tag indicating direct speech), additional work is needed, and we discuss some further limits of this approach. The resulting parser will be used to parse Early English Books Online, a 1.1 billion word corpus whose utility for the study of syntactic change will be greatly increased with the addition of accurate parse trees.

Jiahong Yuan, Neville Ryant, Xingyu Cai et al.

This study reports our efforts to improve automatic recognition of suprasegmentals by fine-tuning wav2vec 2.0 with CTC, a method that has been successful in automatic speech recognition. We demonstrate that the method can improve the state-of-the-art on automatic recognition of syllables, tones, and pitch accents. Utilizing segmental information, by employing tonal finals or tonal syllables as recognition units, can significantly improve Mandarin tone recognition. Language models are helpful when tonal syllables are used as recognition units, but not helpful when tones are recognition units. Finally, Mandarin tone recognition can benefit from English phoneme recognition by combining the two tasks in fine-tuning wav2vec 2.0.

Neville Ryant, Prachi Singh, Venkat Krishnamohan et al.

DIHARD III was the third in a series of speaker diarization challenges intended to improve the robustness of diarization systems to variability in recording equipment, noise conditions, and conversational domain. Speaker diarization was evaluated under two speech activity conditions (diarization from a reference speech activity vs. diarization from scratch) and 11 diverse domains. The domains span a range of recording conditions and interaction types, including read audio-books, meeting speech, clinical interviews, web videos, and, for the first time, conversational telephone speech. A total of 30 organizations (forming 21teams) from industry and academia submitted 499 valid system outputs. The evaluation results indicate that speaker diarization has improved markedly since DIHARD I, particularly for two-party interactions, but that for many domains (e.g., web video) the problem remains far from solved.

Danni Ma, Neville Ryant, Mark Liberman

Pre-trained acoustic representations such as wav2vec and DeCoAR have attained impressive word error rates (WER) for speech recognition benchmarks, particularly when labeled data is limited. But little is known about what phonetic properties these various representations acquire, and how well they encode transferable features of speech. We compare features from two conventional and four pre-trained systems in some simple frame-level phonetic classification tasks, with classifiers trained on features from one version of the TIMIT dataset and tested on features from another. All contextualized representations offered some level of transferability across domains, and models pre-trained on more audio data give better results; but overall, DeCoAR, the system with the simplest architecture, performs best. This type of benchmarking analysis can thus uncover relative strengths of various proposed acoustic representations.

Neville Ryant, Kenneth Church, Christopher Cieri et al.

This paper introduces the third DIHARD challenge, the third in a series of speaker diarization challenges intended to improve the robustness of diarization systems to variation in recording equipment, noise conditions, and conversational domain. The challenge comprises two tracks evaluating diarization performance when starting from a reference speech segmentation (track 1) and diarization from raw audio scratch (track 2). We describe the task, metrics, datasets, and evaluation protocol.

Seth Kulick, Neville Ryant

We investigate the question of whether advances in NLP over the last few years make it possible to vastly increase the size of data usable for research in historical syntax. This brings together many of the usual tools in NLP - word embeddings, tagging, and parsing - in the service of linguistic queries over automatically annotated corpora. We train a part-of-speech (POS) tagger and parser on a corpus of historical English, using ELMo embeddings trained over a billion words of similar text. The evaluation is based on the standard metrics, as well as on the accuracy of the query searches using the parsed data.

Neville Ryant, Kenneth Church, Christopher Cieri et al.

This paper introduces the second DIHARD challenge, the second in a series of speaker diarization challenges intended to improve the robustness of diarization systems to variation in recording equipment, noise conditions, and conversational domain. The challenge comprises four tracks evaluating diarization performance under two input conditions (single channel vs. multi-channel) and two segmentation conditions (diarization from a reference speech segmentation vs. diarization from scratch). In order to prevent participants from overtuning to a particular combination of recording conditions and conversational domain, recordings are drawn from a variety of sources ranging from read audiobooks to meeting speech, to child language acquisition recordings, to dinner parties, to web video. We describe the task and metrics, challenge design, datasets, and baseline systems for speech enhancement, speech activity detection, and diarization.