arXiv subjects

Tomohiro Tanaka

Publications and source records attributed to Tomohiro Tanaka.

At least 37 records · Page 2Linked to original sources

arXiv2023

Exploration of Language Dependency for Japanese Self-Supervised Speech Representation Models

Self-supervised learning (SSL) has been dramatically successful not only in monolingual but also in cross-lingual settings. However, since the two settings have been studied individually in general, there has been little research focusing on how effective a cross-lingual model is in comparison with a monolingual model. In this paper, we investigate this fundamental question empirically with Japanese automatic speech recognition (ASR) tasks. First, we begin by comparing the ASR performance of cross-lingual and monolingual models for two different language tasks while keeping the acoustic domain as identical as possible. Then, we examine how much unlabeled data collected in Japanese is needed to achieve performance comparable to a cross-lingual model pre-trained with tens of thousands of hours of English and/or multilingual data. Finally, we extensively investigate the effectiveness of SSL in Japanese and demonstrate state-of-the-art performance on multiple ASR tasks. Since there is no comprehensive SSL study for Japanese, we hope this study will guide Japanese SSL research.

cs.CL↗

arXiv2023

Leveraging Large Text Corpora for End-to-End Speech Summarization

End-to-end speech summarization (E2E SSum) is a technique to directly generate summary sentences from speech. Compared with the cascade approach, which combines automatic speech recognition (ASR) and text summarization models, the E2E approach is more promising because it mitigates ASR errors, incorporates nonverbal information, and simplifies the overall system. However, since collecting a large amount of paired data (i.e., speech and summary) is difficult, the training data is usually insufficient to train a robust E2E SSum system. In this paper, we present two novel methods that leverage a large amount of external text summarization data for E2E SSum training. The first technique is to utilize a text-to-speech (TTS) system to generate synthesized speech, which is used for E2E SSum training with the text summary. The second is a TTS-free method that directly inputs phoneme sequence instead of synthesized speech to the E2E SSum model. Experiments show that our proposed TTS- and phoneme-based methods improve several metrics on the How2 dataset. In particular, our best system outperforms a previous state-of-the-art one by a large margin (i.e., METEOR score improvements of more than 6 points). To the best of our knowledge, this is the first work to use external language resources for E2E SSum. Moreover, we report a detailed analysis of the How2 dataset to confirm the validity of our proposed E2E SSum system.

cs.CL↗

arXiv2022

Thermal conductivity of a weakly interacting Bose gas in quasi-one-dimension

Transport coefficients are typically divergent for quantum integrable systems in one dimension, such as a Bose gas with a two-body contact interaction. However, when a one-dimensional system is realized by confining bosons into a tight matter waveguide, an effective three-body interaction inevitably arises as leading perturbation to break the integrability. This fact motivates us to study the thermal conductivity of a Bose gas in one dimension with both two-body and three-body interactions. In particular, we evaluate the Kubo formula exactly to the lowest order in perturbation by summing up all contributions that are naively higher orders in perturbation but become comparable in the zero-frequency limit due to the pinch singularity. Consequently, a self-consistent equation for a vertex function is derived, showing that the thermal conductivity in quasi-one-dimension is dominated by the three-body interaction rather than the two-body interaction. Furthermore, the resulting thermal conductivity in the weak-coupling limit proves to be identical to that computed based on the quantum Boltzmann equation and its temperature dependence is numerically determined.

cond-mat.stat-mech↗

arXiv2022

Ladder Siamese Network: a Method and Insights for Multi-level Self-Supervised Learning

Siamese-network-based self-supervised learning (SSL) suffers from slow convergence and instability in training. To alleviate this, we propose a framework to exploit intermediate self-supervisions in each stage of deep nets, called the Ladder Siamese Network. Our self-supervised losses encourage the intermediate layers to be consistent with different data augmentations to single samples, which facilitates training progress and enhances the discriminative ability of the intermediate layers themselves. While some existing work has already utilized multi-level self supervisions in SSL, ours is different in that 1) we reveal its usefulness with non-contrastive Siamese frameworks in both theoretical and empirical viewpoints, and 2) ours improves image-level classification, instance-level detection, and pixel-level segmentation simultaneously. Experiments show that the proposed framework can improve BYOL baselines by 1.0% points in ImageNet linear classification, 1.2% points in COCO detection, and 3.1% points in PASCAL VOC segmentation. In comparison with the state-of-the-art methods, our Ladder-based model achieves competitive and balanced performances in all tested benchmarks without causing large degradation in one.

cs.CV↗

arXiv2022

Bulk viscosity of dual Bose and Fermi gases in one dimension

One-dimensional Bose and Fermi gases with contact interactions are known to exhibit the weak-strong duality, where the equilibrium thermodynamic properties of one system at weak coupling are identical to those of the other system at strong coupling. Here, we show that such duality extends beyond the thermodynamics to the frequency-dependent complex bulk viscosity, which is provided by the contact-contact response function. In particular, we confirm that the bulk viscosities of the Bose and Fermi gases agree in the high-temperature limit, where the systematic expansion in terms of fugacity is available at arbitrary coupling. We also compute their bulk viscosities perturbatively in the weak-coupling limit at arbitrary temperature, which via the duality serve as those of the Fermi and Bose gases in the strong-coupling limit.

cond-mat.quant-gas↗

arXiv2022

Stochastic optimization of a mixed moving average process for controlling non-Markovian streamflow environments

We investigated a cost-constrained static ergodic control problem of the variance of measure-valued affine processes and its application in streamflow management. The controlled system is a jump-driven mixed moving average process that generates realistic subexponential autocorrelation functions, and the static nature of the control originates from a realistic observability assumption in the system. The Markovian lift was effectively used to discretize the system into a finite-dimensional process, which is easier to analyze. The resolution of the problem is based on backward Kolmogorov equations and a quadratic solution ansatz. The control problem has a closed-form solution, and the variance has both strict upper and lower bounds, indicating that the variance cannot take an arbitrary value even when it is subject to a high control cost. The correspondence between the discretized system based on the Markovian lift and the original infinite-dimensional one is discussed. Then, a convergent Markovian lift is presented to approximate the infinite-dimensional system. Finally, the control problem was applied to real cases using available data for a river reach. An extended problem subject to an additional constraint on maintaining the flow variability was also analyzed without significantly degrading the tractability of the proposed framework.

math.OC↗

arXiv2022

Deep versus Wide: An Analysis of Student Architectures for Task-Agnostic Knowledge Distillation of Self-Supervised Speech Models

Self-supervised learning (SSL) is seen as a very promising approach with high performance for several speech downstream tasks. Since the parameters of SSL models are generally so large that training and inference require a lot of memory and computational cost, it is desirable to produce compact SSL models without a significant performance degradation by applying compression methods such as knowledge distillation (KD). Although the KD approach is able to shrink the depth and/or width of SSL model structures, there has been little research on how varying the depth and width impacts the internal representation of the small-footprint model. This paper provides an empirical study that addresses the question. We investigate the performance on SUPERB while varying the structure and KD methods so as to keep the number of parameters constant; this allows us to analyze the contribution of the representation introduced by varying the model architecture. Experiments demonstrate that a certain depth is essential for solving content-oriented tasks (e.g. automatic speech recognition) accurately, whereas a certain width is necessary for achieving high performance on several speaker-oriented tasks (e.g. speaker identification). Based on these observations, we identify, for SUPERB, a more compressed model with better performance than previous studies.

cs.CL↗

arXiv2022

Strategies to Improve Robustness of Target Speech Extraction to Enrollment Variations

Target speech extraction is a technique to extract the target speaker's voice from mixture signals using a pre-recorded enrollment utterance that characterize the voice characteristics of the target speaker. One major difficulty of target speech extraction lies in handling variability in ``intra-speaker'' characteristics, i.e., characteristics mismatch between target speech and an enrollment utterance. While most conventional approaches focus on improving {\it average performance} given a set of enrollment utterances, here we propose to guarantee the {\it worst performance}, which we believe is of great practical importance. In this work, we propose an evaluation metric called worst-enrollment source-to-distortion ratio (SDR) to quantitatively measure the robustness towards enrollment variations. We also introduce a novel training scheme that aims at directly optimizing the worst-case performance by focusing on training with difficult enrollment cases where extraction does not perform well. In addition, we investigate the effectiveness of auxiliary speaker identification loss (SI-loss) as another way to improve robustness over enrollments. Experimental validation reveals the effectiveness of both worst-enrollment target training and SI-loss training to improve robustness against enrollment variations, by increasing speaker discriminability.

eess.AS↗

arXiv2022

Modeling and computation of an integral operator Riccati equation for an infinite-dimensional stochastic differential equation governing streamflow discharge

We propose a linear-quadratic (LQ) control problem of streamflow discharge by optimizing an infinite-dimensional jump-driven stochastic differential equation (SDE). Our SDE is a superposition of Ornstein-Uhlenbeck processes (supOU process), generating a sub-exponential autocorrelation function observed in actual data. The integral operator Riccati equation is heuristically derived to determine the optimal control of the infinite-dimensional system. In addition, its finite-dimensional version is derived with a discretized distribution of the reversion speed and computed by a finite difference scheme. The optimality of the Riccati equation is analyzed by a verification argument. The supOU process is parameterized based on the actual data of a perennial river. The convergence of the numerical scheme is analyzed through computational experiments. Finally, we demonstrate the application of the proposed model to realistic problems along with the Kolmogorov backward equation for the performance evaluation of controls.

math.OC↗

arXiv2021

Utilizing Resource-Rich Language Datasets for End-to-End Scene Text Recognition in Resource-Poor Languages

This paper presents a novel training method for end-to-end scene text recognition. End-to-end scene text recognition offers high recognition accuracy, especially when using the encoder-decoder model based on Transformer. To train a highly accurate end-to-end model, we need to prepare a large image-to-text paired dataset for the target language. However, it is difficult to collect this data, especially for resource-poor languages. To overcome this difficulty, our proposed method utilizes well-prepared large datasets in resource-rich languages such as English, to train the resource-poor encoder-decoder model. Our key idea is to build a model in which the encoder reflects knowledge of multiple languages while the decoder specializes in knowledge of just the resource-poor language. To this end, the proposed method pre-trains the encoder by using a multilingual dataset that combines the resource-poor language's dataset and the resource-rich language's dataset to learn language-invariant knowledge for scene text recognition. The proposed method also pre-trains the decoder by using the resource-poor language's dataset to make the decoder better suited to the resource-poor language. Experiments on Japanese scene text recognition using a small, publicly available dataset demonstrate the effectiveness of the proposed method.

cs.CV↗

arXiv2021

Hierarchical Knowledge Distillation for Dialogue Sequence Labeling

This paper presents a novel knowledge distillation method for dialogue sequence labeling. Dialogue sequence labeling is a supervised learning task that estimates labels for each utterance in the target dialogue document, and is useful for many applications such as dialogue act estimation. Accurate labeling is often realized by a hierarchically-structured large model consisting of utterance-level and dialogue-level networks that capture the contexts within an utterance and between utterances, respectively. However, due to its large model size, such a model cannot be deployed on resource-constrained devices. To overcome this difficulty, we focus on knowledge distillation which trains a small model by distilling the knowledge of a large and high performance teacher model. Our key idea is to distill the knowledge while keeping the complex contexts captured by the teacher model. To this end, the proposed method, hierarchical knowledge distillation, trains the small model by distilling not only the probability distribution of the label classification, but also the knowledge of utterance-level and dialogue-level contexts trained in the teacher model by training the model to mimic the teacher model's output in each level. Experiments on dialogue act estimation and call scene segmentation demonstrate the effectiveness of the proposed method.

cs.CL↗

arXiv2021

End-to-End Rich Transcription-Style Automatic Speech Recognition with Semi-Supervised Learning

We propose a semi-supervised learning method for building end-to-end rich transcription-style automatic speech recognition (RT-ASR) systems from small-scale rich transcription-style and large-scale common transcription-style datasets. In spontaneous speech tasks, various speech phenomena such as fillers, word fragments, laughter and coughs, etc. are often included. While common transcriptions do not give special awareness to these phenomena, rich transcriptions explicitly convert them into special phenomenon tokens as well as textual tokens. In previous studies, the textual and phenomenon tokens were simultaneously estimated in an end-to-end manner. However, it is difficult to build accurate RT-ASR systems because large-scale rich transcription-style datasets are often unavailable. To solve this problem, our training method uses a limited rich transcription-style dataset and common transcription-style dataset simultaneously. The Key process in our semi-supervised learning is to convert the common transcription-style dataset into a pseudo-rich transcription-style dataset. To this end, we introduce style tokens which control phenomenon tokens are generated or not into transformer-based autoregressive modeling. We use this modeling for generating the pseudo-rich transcription-style datasets and for building RT-ASR system from the pseudo and original datasets. Our experiments on spontaneous ASR tasks showed the effectiveness of the proposed method.

cs.CL↗

arXiv2021

Unified Autoregressive Modeling for Joint End-to-End Multi-Talker Overlapped Speech Recognition and Speaker Attribute Estimation

In this paper, we present a novel modeling method for single-channel multi-talker overlapped automatic speech recognition (ASR) systems. Fully neural network based end-to-end models have dramatically improved the performance of multi-taker overlapped ASR tasks. One promising approach for end-to-end modeling is autoregressive modeling with serialized output training in which transcriptions of multiple speakers are recursively generated one after another. This enables us to naturally capture relationships between speakers. However, the conventional modeling method cannot explicitly take into account the speaker attributes of individual utterances such as gender and age information. In fact, the performance deteriorates when each speaker is the same gender or is close in age. To address this problem, we propose unified autoregressive modeling for joint end-to-end multi-talker overlapped ASR and speaker attribute estimation. Our key idea is to handle gender and age estimation tasks within the unified autoregressive modeling. In the proposed method, transformer-based autoregressive model recursively generates not only textual tokens but also attribute tokens of each speaker. This enables us to effectively utilize speaker attributes for improving multi-talker overlapped ASR. Experiments on Japanese multi-talker overlapped ASR tasks demonstrate the effectiveness of the proposed method.

cs.CL↗

arXiv2021

Cross-Modal Transformer-Based Neural Correction Models for Automatic Speech Recognition

We propose a cross-modal transformer-based neural correction models that refines the output of an automatic speech recognition (ASR) system so as to exclude ASR errors. Generally, neural correction models are composed of encoder-decoder networks, which can directly model sequence-to-sequence mapping problems. The most successful method is to use both input speech and its ASR output text as the input contexts for the encoder-decoder networks. However, the conventional method cannot take into account the relationships between these two different modal inputs because the input contexts are separately encoded for each modal. To effectively leverage the correlated information between the two different modal inputs, our proposed models encode two different contexts jointly on the basis of cross-modal self-attention using a transformer. We expect that cross-modal self-attention can effectively capture the relationships between two different modals for refining ASR hypotheses. We also introduce a shallow fusion technique to efficiently integrate the first-pass ASR model and our proposed neural correction model. Experiments on Japanese natural language ASR tasks demonstrated that our proposed models achieve better ASR performance than conventional neural correction models.

cs.CL↗

arXiv2021

Zero-Shot Joint Modeling of Multiple Spoken-Text-Style Conversion Tasks using Switching Tokens

In this paper, we propose a novel spoken-text-style conversion method that can simultaneously execute multiple style conversion modules such as punctuation restoration and disfluency deletion without preparing matched datasets. In practice, transcriptions generated by automatic speech recognition systems are not highly readable because they often include many disfluencies and do not include punctuation marks. To improve their readability, multiple spoken-text-style conversion modules that individually model a single conversion task are cascaded because matched datasets that simultaneously handle multiple conversion tasks are often unavailable. However, the cascading is unstable against the order of tasks because of the chain of conversion errors. Besides, the computation cost of the cascading must be higher than the single conversion. To execute multiple conversion tasks simultaneously without preparing matched datasets, our key idea is to distinguish individual conversion tasks using the on-off switch. In our proposed zero-shot joint modeling, we switch the individual tasks using multiple switching tokens, enabling us to utilize a zero-shot learning approach to executing simultaneous conversions. Our experiments on joint modeling of disfluency deletion and punctuation restoration demonstrate the effectiveness of our method.

cs.CL↗

arXiv2021

Enrollment-less training for personalized voice activity detection

We present a novel personalized voice activity detection (PVAD) learning method that does not require enrollment data during training. PVAD is a task to detect the speech segments of a specific target speaker at the frame level using enrollment speech of the target speaker. Since PVAD must learn speakers' speech variations to clarify the boundary between speakers, studies on PVAD used large-scale datasets that contain many utterances for each speaker. However, the datasets to train a PVAD model are often limited because substantial cost is needed to prepare such a dataset. In addition, we cannot utilize the datasets used to train the standard VAD because they often lack speaker labels. To solve these problems, our key idea is to use one utterance as both a kind of enrollment speech and an input to the PVAD during training, which enables PVAD training without enrollment speech. In our proposed method, called enrollment-less training, we augment one utterance so as to create variability between the input and the enrollment speech while keeping the speaker identity, which avoids the mismatch between training and inference. Our experimental results demonstrate the efficacy of the method.

cs.SD↗

arXiv2021

Context-Free TextSpotter for Real-Time and Mobile End-to-End Text Detection and Recognition

In the deployment of scene-text spotting systems on mobile platforms, lightweight models with low computation are preferable. In concept, end-to-end (E2E) text spotting is suitable for such purposes because it performs text detection and recognition in a single model. However, current state-of-the-art E2E methods rely on heavy feature extractors, recurrent sequence modellings, and complex shape aligners to pursue accuracy, which means their computations are still heavy. We explore the opposite direction: How far can we go without bells and whistles in E2E text spotting? To this end, we propose a text-spotting method that consists of simple convolutions and a few post-processes, named Context-Free TextSpotter. Experiments using standard benchmarks show that Context-Free TextSpotter achieves real-time text spotting on a GPU with only three million parameters, which is the smallest and fastest among existing deep text spotters, with an acceptable transcription quality degradation compared to heavier ones. Further, we demonstrate that our text spotter can run on a smartphone with affordable latency, which is valuable for building stand-alone OCR applications.

cs.CV↗

arXiv2021

Audio-Visual Speech Separation Using Cross-Modal Correspondence Loss

We present an audio-visual speech separation learning method that considers the correspondence between the separated signals and the visual signals to reflect the speech characteristics during training. Audio-visual speech separation is a technique to estimate the individual speech signals from a mixture using the visual signals of the speakers. Conventional studies on audio-visual speech separation mainly train the separation model on the audio-only loss, which reflects the distance between the source signals and the separated signals. However, conventional losses do not reflect the characteristics of the speech signals, including the speaker's characteristics and phonetic information, which leads to distortion or remaining noise. To address this problem, we propose the cross-modal correspondence (CMC) loss, which is based on the cooccurrence of the speech signal and the visual signal. Since the visual signal is not affected by background noise and contains speaker and phonetic information, using the CMC loss enables the audio-visual speech separation model to remove noise while preserving the speech characteristics. Experimental results demonstrate that the proposed method learns the cooccurrence on the basis of CMC loss, which improves separation performance.

cs.SD↗