Panayiotis Georgiou

Dominik Wagner, Alexander Churchill, Siddharth Sigtia et al.

Interactions with virtual assistants typically start with a predefined trigger phrase followed by the user command. To make interactions with the assistant more intuitive, we explore whether it is feasible to drop the requirement that users must begin each command with a trigger phrase. We explore this task in three ways: First, we train classifiers using only acoustic information obtained from the audio waveform. Second, we take the decoder outputs of an automatic speech recognition (ASR) system, such as 1-best hypotheses, as input features to a large language model (LLM). Finally, we explore a multimodal system that combines acoustic and lexical features, as well as ASR decoder signals in an LLM. Using multimodal information yields relative equal-error-rate improvements over text-only and audio-only models of up to 39% and 61%. Increasing the size of the LLM and training with low-rank adaption leads to further relative EER reductions of up to 18% on our dataset.

Dominik Wagner, Alexander Churchill, Siddharth Sigtia et al.

Interactions with virtual assistants typically start with a trigger phrase followed by a command. In this work, we explore the possibility of making these interactions more natural by eliminating the need for a trigger phrase. Our goal is to determine whether a user addressed the virtual assistant based on signals obtained from the streaming audio recorded by the device microphone. We address this task by combining 1-best hypotheses and decoder signals from an automatic speech recognition system with acoustic representations from an audio encoder as input features to a large language model (LLM). In particular, we are interested in data and resource efficient systems that require only a small amount of training data and can operate in scenarios with only a single frozen LLM available on a device. For this reason, our model is trained on 80k or less examples of multimodal data using a combination of low-rank adaptation and prefix tuning. We compare the proposed system to unimodal baselines and show that the multimodal approach achieves lower equal-error-rates (EERs), while using only a fraction of the training data. We also show that low-dimensional specialized audio representations lead to lower EERs than high-dimensional general audio representations.

Vin Sachidananda, Shao-Yen Tseng, Erik Marchi et al.

Deriving multimodal representations of audio and lexical inputs is a central problem in Natural Language Understanding (NLU). In this paper, we present Contrastive Aligned Audio-Language Multirate and Multimodal Representations (CALM), an approach for learning multimodal representations using contrastive and multirate information inherent in audio and lexical inputs. The proposed model aligns acoustic and lexical information in the input embedding space of a pretrained language-only contextual embedding model. By aligning audio representations to pretrained language representations and utilizing contrastive information between acoustic inputs, CALM is able to bootstrap audio embedding competitive with existing audio representation models in only a few hours of training time. Operationally, audio spectrograms are processed using linearized patches through a Spectral Transformer (SpecTran) which is trained using a Contrastive Audio-Language Pretraining objective to align audio and language from similar queries. Subsequently, the derived acoustic and lexical tokens representations are input into a multimodal transformer to incorporate utterance level context and derive the proposed CALM representations. We show that these pretrained embeddings can subsequently be used in multimodal supervised tasks and demonstrate the benefits of the proposed pretraining steps in terms of the alignment of the two embedding spaces and the multirate nature of the pretraining. Our system shows 10-25\% improvement over existing emotion recognition systems including state-of-the-art three-modality systems under various evaluation objectives.

Vikramjit Mitra, Zifang Huang, Colin Lea et al.

Dysfluencies and variations in speech pronunciation can severely degrade speech recognition performance, and for many individuals with moderate-to-severe speech disorders, voice operated systems do not work. Current speech recognition systems are trained primarily with data from fluent speakers and as a consequence do not generalize well to speech with dysfluencies such as sound or word repetitions, sound prolongations, or audible blocks. The focus of this work is on quantitative analysis of a consumer speech recognition system on individuals who stutter and production-oriented approaches for improving performance for common voice assistant tasks (i.e., "what is the weather?"). At baseline, this system introduces a significant number of insertion and substitution errors resulting in intended speech Word Error Rates (isWER) that are 13.64\% worse (absolute) for individuals with fluency disorders. We show that by simply tuning the decoding parameters in an existing hybrid speech recognition system one can improve isWER by 24\% (relative) for individuals with fluency disorders. Tuning these parameters translates to 3.6\% better domain recognition and 1.7\% better intent recognition relative to the default setup for the 18 study participants across all stuttering severities.

Haoqi Li, Brian Baucom, Shrikanth Narayanan et al.

Speech encodes a wealth of information related to human behavior and has been used in a variety of automated behavior recognition tasks. However, extracting behavioral information from speech remains challenging including due to inadequate training data resources stemming from the often low occurrence frequencies of specific behavioral patterns. Moreover, supervised behavioral modeling typically relies on domain-specific construct definitions and corresponding manually-annotated data, rendering generalizing across domains challenging. In this paper, we exploit the stationary properties of human behavior within an interaction and present a representation learning method to capture behavioral information from speech in an unsupervised way. We hypothesize that nearby segments of speech share the same behavioral context and hence map onto similar underlying behavioral representations. We present an encoder-decoder based Deep Contextualized Network (DCN) as well as a Triplet-Enhanced DCN (TE-DCN) framework to capture the behavioral context and derive a manifold representation, where speech frames with similar behaviors are closer while frames of different behaviors maintain larger distances. The models are trained on movie audio data and validated on diverse domains including on a couples therapy corpus and other publicly collected data (e.g., stand-up comedy). With encouraging results, our proposed framework shows the feasibility of unsupervised learning within cross-domain behavioral modeling.

Nikolaos Flemotomos, Victor R. Martinez, Zhuohao Chen et al.

With the growing prevalence of psychological interventions, it is vital to have measures which rate the effectiveness of psychological care to assist in training, supervision, and quality assurance of services. Traditionally, quality assessment is addressed by human raters who evaluate recorded sessions along specific dimensions, often codified through constructs relevant to the approach and domain. This is however a cost-prohibitive and time-consuming method that leads to poor feasibility and limited use in real-world settings. To facilitate this process, we have developed an automated competency rating tool able to process the raw recorded audio of a session, analyzing who spoke when, what they said, and how the health professional used language to provide therapy. Focusing on a use case of a specific type of psychotherapy called Motivational Interviewing, our system gives comprehensive feedback to the therapist, including information about the dynamics of the session (e.g., therapist's vs. client's talking time), low-level psychological language descriptors (e.g., type of questions asked), as well as other high-level behavioral constructs (e.g., the extent to which the therapist understands the clients' perspective). We describe our platform and its performance using a dataset of more than 5,000 recordings drawn from its deployment in a real-world clinical setting used to assist training of new therapists. Widespread use of automated psychotherapy rating tools may augment experts' capabilities by providing an avenue for more effective training and skill improvement, eventually leading to more positive clinical outcomes.

Prashanth Gurunath Shivakumar, Panayiotis Georgiou, Shrikanth Narayanan

Word vector representations enable machines to encode human language for spoken language understanding and processing. Confusion2vec, motivated from human speech production and perception, is a word vector representation which encodes ambiguities present in human spoken language in addition to semantics and syntactic information. Confusion2vec provides a robust spoken language representation by considering inherent human language ambiguities. In this paper, we propose a novel word vector space estimation by unsupervised learning on lattices output by an automatic speech recognition (ASR) system. We encode each word in confusion2vec vector space by its constituent subword character n-grams. We show the subword encoding helps better represent the acoustic perceptual ambiguities in human spoken language via information modeled on lattice structured ASR output. The usefulness of the proposed Confusion2vec representation is evaluated using semantic, syntactic and acoustic analogy and word similarity tasks. We also show the benefits of subword modeling for acoustic ambiguity representation on the task of spoken language intent detection. The results significantly outperform existing word vector representations when evaluated on erroneous ASR outputs. We demonstrate that Confusion2vec subword modeling eliminates the need for retraining/adapting the natural language understanding models on ASR transcripts.

Tae Jin Park, Kyu J. Han, Jing Huang et al.

This work presents a novel approach for speaker diarization to leverage lexical information provided by automatic speech recognition. We propose a speaker diarization system that can incorporate word-level speaker turn probabilities with speaker embeddings into a speaker clustering process to improve the overall diarization accuracy. To integrate lexical and acoustic information in a comprehensive way during clustering, we introduce an adjacency matrix integration for spectral clustering. Since words and word boundary information for word-level speaker turn probability estimation are provided by a speech recognition system, our proposed method works without any human intervention for manual transcriptions. We show that the proposed method improves diarization performance on various evaluation datasets compared to the baseline diarization system using acoustic information only in speaker embeddings.

Sandeep Nallan Chakravarthula, Brian Baucom, Shrikanth Narayanan et al.

The task of quantifying human behavior by observing interaction cues is an important and useful one across a range of domains in psychological research and practice. Machine learning-based approaches typically perform this task by first estimating behavior based on cues within an observation window, such as a fixed number of words, and then aggregating the behavior over all the windows in that interaction. The length of this window directly impacts the accuracy of estimation by controlling the amount of information being used. The exact link between window length and accuracy, however, has not been well studied, especially in spoken language. In this paper, we investigate this link and present an analysis framework that determines appropriate window lengths for the task of behavior estimation. Our proposed framework utilizes a two-pronged evaluation approach: (a) extrinsic similarity between machine predictions and human expert annotations, and (b) intrinsic consistency between intra-machine and intra-human behavior relations. We apply our analysis to real-life conversations that are annotated for a large and diverse set of behavior codes and examine the relation between the nature of a behavior and how long it should be observed. We find that behaviors describing negative and positive affect can be accurately estimated from short to medium-length expressions whereas behaviors related to problem-solving and dysphoria require much longer observations and are difficult to quantify from language alone. These findings are found to be generally consistent across different behavior modeling approaches.

Nikolaos Flemotomos, Panayiotis Georgiou, Shrikanth Narayanan

Speaker diarization relies on the assumption that speech segments corresponding to a particular speaker are concentrated in a specific region of the speaker space; a region which represents that speaker's identity. These identities are not known a priori, so a clustering algorithm is typically employed, which is traditionally based solely on audio. Under noisy conditions, however, such an approach poses the risk of generating unreliable speaker clusters. In this work we aim to utilize linguistic information as a supplemental modality to identify the various speakers in a more robust way. We are focused on conversational scenarios where the speakers assume distinct roles and are expected to follow different linguistic patterns. This distinct linguistic variability can be exploited to help us construct the speaker identities. That way, we are able to boost the diarization performance by converting the clustering task to a classification one. The proposed method is applied in real-world dyadic psychotherapy interactions between a provider and a patient and demonstrated to show improved results.

Haoqi Li, Ming Tu, Jing Huang et al.

Representation learning for speech emotion recognition is challenging due to labeled data sparsity issue and lack of gold standard references. In addition, there is much variability from input speech signals, human subjective perception of the signals and emotion label ambiguity. In this paper, we propose a machine learning framework to obtain speech emotion representations by limiting the effect of speaker variability in the speech signals. Specifically, we propose to disentangle the speaker characteristics from emotion through an adversarial training network in order to better represent emotion. Our method combines the gradient reversal technique with an entropy loss function to remove such speaker information. Our approach is evaluated on both IEMOCAP and CMU-MOSEI datasets. We show that our method improves speech emotion classification and increases generalization to unseen speakers.

Prashanth Gurunath Shivakumar, Naveen Kumar, Panayiotis Georgiou et al.

Spoken Language Understanding (SLU) typically comprises of an automatic speech recognition (ASR) followed by a natural language understanding (NLU) module. The two modules process signals in a blocking sequential fashion, i.e., the NLU often has to wait for the ASR to finish processing on an utterance basis, potentially leading to high latencies that render the spoken interaction less natural. In this paper, we propose recurrent neural network (RNN) based incremental processing towards the SLU task of intent detection. The proposed methodology offers lower latencies than a typical SLU system, without any significant reduction in system accuracy. We introduce and analyze different recurrent neural network architectures for incremental and online processing of the ASR transcripts and compare it to the existing offline systems. A lexical End-of-Sentence (EOS) detector is proposed for segmenting the stream of transcript into sentences for intent classification. Intent detection experiments are conducted on benchmark ATIS, Snips and Facebook's multilingual task oriented dialog datasets modified to emulate a continuous incremental stream of words with no utterance demarcation. We also analyze the prospects of early intent detection, before EOS, with our proposed system.

Haoqi Li, Brian Baucom, Panayiotis Georgiou

Human behavior refers to the way humans act and interact. Understanding human behavior is a cornerstone of observational practice, especially in psychotherapy. An important cue of behavior analysis is the dynamical changes of emotions during the conversation. Domain experts integrate emotional information in a highly nonlinear manner, thus, it is challenging to explicitly quantify the relationship between emotions and behaviors. In this work, we employ deep transfer learning to analyze their inferential capacity and contextual importance. We first train a network to quantify emotions from acoustic signals and then use information from the emotion recognition network as features for behavior recognition. We treat this emotion-related information as behavioral primitives and further train higher level layers towards behavior quantification. Through our analysis, we find that emotion-related information is an important cue for behavior recognition. Further, we investigate the importance of emotional-context in the expression of behavior by constraining (or not) the neural networks' contextual view of the data. This demonstrates that the sequence of emotions is critical in behavior expression. To achieve these frameworks we employ hybrid architectures of convolutional networks and recurrent networks to extract emotion-related behavior primitives and facilitate automatic behavior recognition from speech.

Shao-Yen Tseng, Panayiotis Georgiou, Shrikanth Narayanan

Word embeddings such as ELMo have recently been shown to model word semantics with greater efficacy through contextualized learning on large-scale language corpora, resulting in significant improvement in state of the art across many natural language tasks. In this work we integrate acoustic information into contextualized lexical embeddings through the addition of multimodal inputs to a pretrained bidirectional language model. The language model is trained on spoken language that includes text and audio modalities. The resulting representations from this model are multimodal and contain paralinguistic information which can modify word meanings and provide affective information. We show that these multimodal embeddings can be used to improve over previous state of the art multimodal models in emotion recognition on the CMU-MOSEI dataset.

Prashanth Gurunath Shivakumar, Shao-Yen Tseng, Panayiotis Georgiou et al.

Most current language modeling techniques only exploit co-occurrence, semantic and syntactic information from the sequence of words. However, a range of information such as the state of the speaker and dynamics of the interaction might be useful. In this work we derive motivation from psycholinguistics and propose the addition of behavioral information into the context of language modeling. We propose the augmentation of language models with an additional module which analyzes the behavioral state of the current context. This behavioral information is used to gate the outputs of the language model before the final word prediction output. We show that the addition of behavioral context in language models achieves lower perplexities on behavior-rich datasets. We also confirm the validity of the proposed models on a variety of model architectures and improve on previous state-of-the-art models with generic domain Penn Treebank Corpus.

Sandeep Nallan Chakravarthula, Haoqi Li, Shao-Yen Tseng et al.

Cancer impacts the quality of life of those diagnosed as well as their spouse caregivers, in addition to potentially influencing their day-to-day behaviors. There is evidence that effective communication between spouses can improve well-being related to cancer but it is difficult to efficiently evaluate the quality of daily life interactions using manual annotation frameworks. Automated recognition of behaviors based on the interaction cues of speakers can help analyze interactions in such couples and identify behaviors which are beneficial for effective communication. In this paper, we present and detail a dataset of dyadic interactions in 85 real-life cancer-afflicted couples and a set of observational behavior codes pertaining to interpersonal communication attributes. We describe and employ neural network-based systems for classifying these behaviors based on turn-level acoustic and lexical speech patterns. Furthermore, we investigate the effect of controlling for factors such as gender, patient/caregiver role and conversation content on behavior classification. Analysis of our preliminary results indicates the challenges in this task due to the nature of the targeted behaviors and suggests that techniques incorporating contextual processing might be better suited to tackle this problem.

Md Nasir, Sandeep Nallan Chakravarthula, Brian Baucom et al.

Linguistic coordination is a well-established phenomenon in spoken conversations and often associated with positive social behaviors and outcomes. While there have been many attempts to measure lexical coordination or entrainment in literature, only a few have explored coordination in syntactic or semantic space. In this work, we attempt to combine these different aspects of coordination into a single measure by leveraging distances in a neural word representation space. In particular, we adopt the recently proposed Word Mover's Distance with word2vec embeddings and extend it to measure the dissimilarity in language used in multiple consecutive speaker turns. To validate our approach, we apply this measure for two case studies in the clinical psychology domain. We find that our proposed measure is correlated with the therapist's empathy towards their patient in Motivational Interviewing and with affective behaviors in Couples Therapy. In both case studies, our proposed metric exhibits higher correlation than previously proposed measures. When applied to the couples with relationship improvement, we also notice a significant decrease in the proposed measure over the course of therapy, indicating higher linguistic coordination.

Prashanth Gurunath Shivakumar, Mu Yang, Panayiotis Georgiou

Decoding speaker's intent is a crucial part of spoken language understanding (SLU). The presence of noise or errors in the text transcriptions, in real life scenarios make the task more challenging. In this paper, we address the spoken language intent detection under noisy conditions imposed by automatic speech recognition (ASR) systems. We propose to employ confusion2vec word feature representation to compensate for the errors made by ASR and to increase the robustness of the SLU system. The confusion2vec, motivated from human speech production and perception, models acoustic relationships between words in addition to the semantic and syntactic relations of words in human language. We hypothesize that ASR often makes errors relating to acoustically similar words, and the confusion2vec with inherent model of acoustic relationships between words is able to compensate for the errors. We demonstrate through experiments on the ATIS benchmark dataset, the robustness of the proposed model to achieve state-of-the-art results under noisy ASR conditions. Our system reduces classification error rate (CER) by 20.84% and improves robustness by 37.48% (lower CER degradation) relative to the previous state-of-the-art going from clean to noisy transcripts. Improvements are also demonstrated when training the intent detection models on noisy transcripts.

Tae Jin Park, Kyu Han, Ian Lane et al.

This work presents a novel approach to leverage lexical information for speaker diarization. We introduce a speaker diarization system that can directly integrate lexical as well as acoustic information into a speaker clustering process. Thus, we propose an adjacency matrix integration technique to integrate word level speaker turn probabilities with speaker embeddings in a comprehensive way. Our proposed method works without any reference transcript. Words, and word boundary information are provided by an ASR system. We show that our proposed method improves a baseline speaker diarization system solely based on speaker embeddings, achieving a meaningful improvement on the CALLHOME American English Speech dataset.

Prashanth Gurunath Shivakumar, Panayiotis Georgiou

Word vector representations are a crucial part of Natural Language Processing (NLP) and Human Computer Interaction. In this paper, we propose a novel word vector representation, Confusion2Vec, motivated from the human speech production and perception that encodes representational ambiguity. Humans employ both acoustic similarity cues and contextual cues to decode information and we focus on a model that incorporates both sources of information. The representational ambiguity of acoustics, which manifests itself in word confusions, is often resolved by both humans and machines through contextual cues. A range of representational ambiguities can emerge in various domains further to acoustic perception, such as morphological transformations, paraphrasing for NLP tasks like machine translation etc. In this work, we present a case study in application to Automatic Speech Recognition (ASR), where the word confusions are related to acoustic similarity. We present several techniques to train an acoustic perceptual similarity representation ambiguity. We term this Confusion2Vec and learn on unsupervised-generated data from ASR confusion networks or lattice-like structures. Appropriate evaluations for the Confusion2Vec are formulated for gauging acoustic similarity in addition to semantic-syntactic and word similarity evaluations. The Confusion2Vec is able to model word confusions efficiently, without compromising on the semantic-syntactic word relations, thus effectively enriching the word vector space with extra task relevant ambiguity information. We provide an intuitive exploration of the 2-dimensional Confusion2Vec space using Principal Component Analysis of the embedding and relate to semantic, syntactic and acoustic relationships. The potential of Confusion2Vec in the utilization of uncertainty present in lattices is demonstrated through small examples relating to ASR error correction.

James Gibson, David C. Atkins, Torrey Creed et al.

We propose a methodology for estimating human behaviors in psychotherapy sessions using mutli-label and multi-task learning paradigms. We discuss the problem of behavioral coding in which data of human interactions is the annotated with labels to describe relevant human behaviors of interest. We describe two related, yet distinct, corpora consisting of therapist client interactions in psychotherapy sessions. We experimentally compare the proposed learning approaches for estimating behaviors of interest in these datasets. Specifically, we compare single and multiple label learning approaches, single and multiple task learning approaches, and evaluate the performance of these approaches when incorporating turn context. We demonstrate the prediction performance gains which can be achieved by using the proposed paradigms and discuss the insights these models provide into these complex interactions.

Shao-Yen Tseng, Brian Baucom, Panayiotis Georgiou

Unsupervised learning has been an attractive method for easily deriving meaningful data representations from vast amounts of unlabeled data. These representations, or embeddings, often yield superior results in many tasks, whether used directly or as features in subsequent training stages. However, the quality of the embeddings is highly dependent on the assumed knowledge in the unlabeled data and how the system extracts information without supervision. Domain portability is also very limited in unsupervised learning, often requiring re-training on other in-domain corpora to achieve robustness. In this work we present a multitask paradigm for unsupervised contextual learning of behavioral interactions which addresses unsupervised domain adaption. We introduce an online multitask objective into unsupervised learning and show that sentence embeddings generated through this process increases performance of affective tasks.

Tae Jin Park, Panayiotis Georgiou

While there has been substantial amount of work in speaker diarization recently, there are few efforts in jointly employing lexical and acoustic information for speaker segmentation. Towards that, we investigate a speaker diarization system using a sequence-to-sequence neural network trained on both lexical and acoustic features. We also propose a loss function that allows for selecting not only the speaker change points but also the best speaker at any time by allowing for different speaker groupings. We incorporate Mel Frequency Cepstral Coefficients (MFCC) as an acoustic feature alongside lexical information that are obtained from conversations from the Fisher dataset. Thus, we show that acoustics provide complementary information to the lexical modality. The experimental results show that sequence-to-sequence system trained on both word sequences and MFCC can improve on speaker diarization result compared to the system that only relies on lexical modality or the baseline MFCC-based system. In addition, we test the performance of our proposed method with Automatic Speech Recognition (ASR) transcripts. While the performance on ASR transcripts drops, the Diarization Error Rate (DER) of our proposed method still outperforms the traditional method based on Bayesian Information Criterion (BIC).

Sandeep Nallan Chakravarthula, Brian Baucom, Panayiotis Georgiou

Dyadic interactions among humans are marked by speakers continuously influencing and reacting to each other in terms of responses and behaviors, among others. Understanding how interpersonal dynamics affect behavior is important for successful treatment in psychotherapy domains. Traditional schemes that automatically identify behavior for this purpose have often looked at only the target speaker. In this work, we propose a Markov model of how a target speaker's behavior is influenced by their own past behavior as well as their perception of their partner's behavior, based on lexical features. Apart from incorporating additional potentially useful information, our model can also control the degree to which the partner affects the target speaker. We evaluate our proposed model on the task of classifying Negative behavior in Couples Therapy and show that it is more accurate than the single-speaker model. Furthermore, we investigate the degree to which the optimal influence relates to how well a couple does on the long-term, via relating to relationship outcomes

Prashanth Gurunath Shivakumar, Panayiotis Georgiou

Children speech recognition is challenging mainly due to the inherent high variability in children's physical and articulatory characteristics and expressions. This variability manifests in both acoustic constructs and linguistic usage due to the rapidly changing developmental stage in children's life. Part of the challenge is due to the lack of large amounts of available children speech data for efficient modeling. This work attempts to address the key challenges using transfer learning from adult's models to children's models in a Deep Neural Network (DNN) framework for children's Automatic Speech Recognition (ASR) task evaluating on multiple children's speech corpora with a large vocabulary. The paper presents a systematic and an extensive analysis of the proposed transfer learning technique considering the key factors affecting children's speech recognition from prior literature. Evaluations are presented on (i) comparisons of earlier GMM-HMM and the newer DNN Models, (ii) effectiveness of standard adaptation techniques versus transfer learning, (iii) various adaptation configurations in tackling the variabilities present in children speech, in terms of (a) acoustic spectral variability, and (b) pronunciation variability and linguistic constraints. Our Analysis spans over (i) number of DNN model parameters (for adaptation), (ii) amount of adaptation data, (iii) ages of children, (iv) age dependent-independent adaptation. Finally, we provide Recommendations on (i) the favorable strategies over various aforementioned - analyzed parameters, and (ii) potential future research directions and relevant challenges/problems persisting in DNN based ASR for children's speech.

Md Nasir, Brian Baucom, Shrikanth Narayanan et al.

Entrainment is a known adaptation mechanism that causes interaction participants to adapt or synchronize their acoustic characteristics. Understanding how interlocutors tend to adapt to each other's speaking style through entrainment involves measuring a range of acoustic features and comparing those via multiple signal comparison methods. In this work, we present a turn-level distance measure obtained in an unsupervised manner using a Deep Neural Network (DNN) model, which we call Neural Entrainment Distance (NED). This metric establishes a framework that learns an embedding from the population-wide entrainment in an unlabeled training corpus. We use the framework for a set of acoustic features and validate the measure experimentally by showing its efficacy in distinguishing real conversations from fake ones created by randomly shuffling speaker turns. Moreover, we show real world evidence of the validity of the proposed measure. We find that high value of NED is associated with high ratings of emotional bond in suicide assessment interviews, which is consistent with prior studies.

Arindam Jati, Panayiotis Georgiou

Learning speaker-specific features is vital in many applications like speaker recognition, diarization and speech recognition. This paper provides a novel approach, we term Neural Predictive Coding (NPC), to learn speaker-specific characteristics in a completely unsupervised manner from large amounts of unlabeled training data that even contain many non-speech events and multi-speaker audio streams. The NPC framework exploits the proposed short-term active-speaker stationarity hypothesis which assumes two temporally-close short speech segments belong to the same speaker, and thus a common representation that can encode the commonalities of both the segments, should capture the vocal characteristics of that speaker. We train a convolutional deep siamese network to produce "speaker embeddings" by learning to separate `same' vs `different' speaker pairs which are generated from an unlabeled data of audio streams. Two sets of experiments are done in different scenarios to evaluate the strength of NPC embeddings and compare with state-of-the-art in-domain supervised methods. First, two speaker identification experiments with different context lengths are performed in a scenario with comparatively limited within-speaker channel variability. NPC embeddings are found to perform the best at short duration experiment, and they provide complementary information to i-vectors for full utterance experiments. Second, a large scale speaker verification task having a wide range of within-speaker channel variability is adopted as an upper-bound experiment where comparisons are drawn with in-domain supervised methods.

Prashanth Gurunath Shivakumar, Haoqi Li, Kevin Knight et al.

Automatic speech recognition (ASR) systems often make unrecoverable errors due to subsystem pruning (acoustic, language and pronunciation models); for example pruning words due to acoustics using short-term context, prior to rescoring with long-term context based on linguistics. In this work we model ASR as a phrase-based noisy transformation channel and propose an error correction system that can learn from the aggregate errors of all the independent modules constituting the ASR and attempt to invert those. The proposed system can exploit long-term context using a neural network language model and can better choose between existing ASR output possibilities as well as re-introduce previously pruned or unseen (out-of-vocabulary) phrases. It provides corrections under poorly performing ASR conditions without degrading any accurate transcriptions; such corrections are greater on top of out-of-domain and mismatched data ASR. Our system consistently provides improvements over the baseline ASR, even when baseline is further optimized through recurrent neural network language model rescoring. This demonstrates that any ASR improvements can be exploited independently and that our proposed system can potentially still provide benefits on highly optimized ASR. Finally, we present an extensive analysis of the type of errors corrected by our system.

Haoqi Li, Brian Baucom, Panayiotis Georgiou

Behavioral annotation using signal processing and machine learning is highly dependent on training data and manual annotations of behavioral labels. Previous studies have shown that speech information encodes significant behavioral information and be used in a variety of automated behavior recognition tasks. However, extracting behavior information from speech is still a difficult task due to the sparseness of training data coupled with the complex, high-dimensionality of speech, and the complex and multiple information streams it encodes. In this work we exploit the slow varying properties of human behavior. We hypothesize that nearby segments of speech share the same behavioral context and hence share a similar underlying representation in a latent space. Specifically, we propose a Deep Neural Network (DNN) model to connect behavioral context and derive the behavioral manifold in an unsupervised manner. We evaluate the proposed manifold in the couples therapy domain and also provide examples from publicly available data (e.g. stand-up comedy). We further investigate training within the couples' therapy domain and from movie data. The results are extremely encouraging and promise improved behavioral quantification in an unsupervised manner and warrants further investigation in a range of applications.

Haoqi Li, Brian Baucom, Panayiotis Georgiou

Observational studies are based on accurate assessment of human state. A behavior recognition system that models interlocutors' state in real-time can significantly aid the mental health domain. However, behavior recognition from speech remains a challenging task since it is difficult to find generalizable and representative features because of noisy and high-dimensional data, especially when data is limited and annotated coarsely and subjectively. Deep Neural Networks (DNN) have shown promise in a wide range of machine learning tasks, but for Behavioral Signal Processing (BSP) tasks their application has been constrained due to limited quantity of data. We propose a Sparsely-Connected and Disjointly-Trained DNN (SD-DNN) framework to deal with limited data. First, we break the acoustic feature set into subsets and train multiple distinct classifiers. Then, the hidden layers of these classifiers become parts of a deeper network that integrates all feature streams. The overall system allows for full connectivity while limiting the number of parameters trained at any time and allows convergence possible with even limited data. We present results on multiple behavior codes in the couples' therapy domain and demonstrate the benefits in behavior classification accuracy. We also show the viability of this system towards live behavior annotations.