Multi-Modal automatic speech recognition (ASR) techniques aim to leverage additional modalities to improve the performance of speech recognition systems. While existing approaches primarily focus on video or contextual information, the utilization of extra supplementary textual information has been overlooked. Recognizing the abundance of online conference videos with slides, which provide rich domain-specific information in the form of text and images, we release SlideSpeech, a large-scale audio-visual corpus enriched with slides. The corpus contains 1,705 videos, 1,000+ hours, with 473 hours of high-quality transcribed speech. Moreover, the corpus contains a significant amount of real-time synchronized slides. In this work, we present the pipeline for constructing the corpus and propose baseline methods for utilizing text information in the visual slide context. Through the application of keyword extraction and contextual ASR methods in the benchmark system, we demonstrate the potential of improving speech recognition performance by incorporating textual information from supplementary video slides.
Hotword customization is one of the important issues remained in ASR field - it is of value to enable users of ASR systems to customize names of entities, persons and other phrases. The past few years have seen both implicit and explicit modeling strategies for ASR contextualization developed. While these approaches have performed adequately, they still exhibit certain shortcomings such as instability in effectiveness. In this paper we propose Semantic-augmented Contextual-Paraformer (SeACo-Paraformer) a novel NAR based ASR system with flexible and effective hotword customization ability. It combines the accuracy of the AED-based model, the efficiency of the NAR model, and the excellent performance in contextualization. In 50,000 hours industrial big data experiments, our proposed model outperforms strong baselines in customization and general ASR tasks. Besides, we explore an efficient way to filter large scale incoming hotwords for further improvement. The source codes and industrial models proposed and compared are all opened as well as two hotword test sets.
The exponential growth of data, alongside advancements in model structures and loss functions, has necessitated the enhancement of image retrieval systems through the utilization of new models with superior feature embeddings. However, the expensive process of updating the old retrieval database by replacing embeddings poses a challenge. As a solution, backward-compatible training can be employed to avoid the necessity of updating old retrieval datasets. While previous methods achieved backward compatibility by aligning prototypes of the old model, they often overlooked the distribution of the old features, thus limiting their effectiveness when the old model's low quality leads to a weakly discriminative feature distribution. On the other hand, instance-based methods like L2 regression take into account the distribution of old features but impose strong constraints on the performance of the new model itself. In this paper, we propose MixBCT, a simple yet highly effective backward-compatible training method that serves as a unified framework for old models of varying qualities. Specifically, we summarize four constraints that are essential for ensuring backward compatibility in an ideal scenario, and we construct a single loss function to facilitate backward-compatible training. Our approach adaptively adjusts the constraint domain for new features based on the distribution of the old embeddings. We conducted extensive experiments on the large-scale face recognition datasets MS1Mv3 and IJB-C to verify the effectiveness of our method. The experimental results clearly demonstrate its superiority over previous methods. Code is available at https://github.com/yuleung/MixBCT
The Audio-Visual Speaker Extraction (AVSE) algorithm employs parallel video recording to leverage two visual cues, namely speaker identity and synchronization, to enhance performance compared to audio-only algorithms. However, the visual front-end in AVSE is often derived from a pre-trained model or end-to-end trained, making it unclear which visual cue contributes more to the speaker extraction performance. This raises the question of how to better utilize visual cues. To address this issue, we propose two training strategies that decouple the learning of the two visual cues. Our experimental results demonstrate that both visual cues are useful, with the synchronization cue having a higher impact. We introduce a more explainable model, the Decoupled Audio-Visual Speaker Extraction (DAVSE) model, which leverages both visual cues.
In recent years, the joint training of speech enhancement front-end and automatic speech recognition (ASR) back-end has been widely used to improve the robustness of ASR systems. Traditional joint training methods only use enhanced speech as input for the backend. However, it is difficult for speech enhancement systems to directly separate speech from input due to the diverse types of noise with different intensities. Furthermore, speech distortion and residual noise are often observed in enhanced speech, and the distortion of speech and noise is different. Most existing methods focus on fusing enhanced and noisy features to address this issue. In this paper, we propose a dual-stream spectrogram refine network to simultaneously refine the speech and noise and decouple the noise from the noisy input. Our proposed method can achieve better performance with a relative 8.6% CER reduction.
Estimating confidence scores for recognition results is a classic task in ASR field and of vital importance for kinds of downstream tasks and training strategies. Previous end-to-end~(E2E) based confidence estimation models (CEM) predict score sequences of equal length with input transcriptions, leading to unreliable estimation when deletion and insertion errors occur. In this paper we proposed CIF-Aligned confidence estimation model (CA-CEM) to achieve accurate and reliable confidence estimation based on novel non-autoregressive E2E ASR model - Paraformer. CA-CEM utilizes the modeling character of continuous integrate-and-fire (CIF) mechanism to generate token-synchronous acoustic embedding, which solves the estimation failure issue above. We measure the quality of estimation with AUC and RMSE in token level and ECE-U - a proposed metrics in utterance level. CA-CEM gains 24% and 19% relative reduction on ECE-U and also better AUC and RMSE on two test sets. Furthermore, we conduct analysis to explore the potential of CEM for different ASR related usage.
The recently proposed serialized output training (SOT) simplifies multi-talker automatic speech recognition (ASR) by generating speaker transcriptions separated by a special token. However, frequent speaker changes can make speaker change prediction difficult. To address this, we propose boundary-aware serialized output training (BA-SOT), which explicitly incorporates boundary knowledge into the decoder via a speaker change detection task and boundary constraint loss. We also introduce a two-stage connectionist temporal classification (CTC) strategy that incorporates token-level SOT CTC to restore temporal context information. Besides typical character error rate (CER), we introduce utterance-dependent character error rate (UD-CER) to further measure the precision of speaker change prediction. Compared to original SOT, BA-SOT reduces CER/UD-CER by 5.1%/14.0%, and leveraging a pre-trained ASR model for BA-SOT model initialization further reduces CER/UD-CER by 8.4%/19.9%.
Recently, speaker-attributed automatic speech recognition (SA-ASR) has attracted a wide attention, which aims at answering the question ``who spoke what''. Different from modular systems, end-to-end (E2E) SA-ASR minimizes the speaker-dependent recognition errors directly and shows a promising applicability. In this paper, we propose a context-aware SA-ASR (CASA-ASR) model by enhancing the contextual modeling ability of E2E SA-ASR. Specifically, in CASA-ASR, a contextual text encoder is involved to aggregate the semantic information of the whole utterance, and a context-dependent scorer is employed to model the speaker discriminability by contrasting with speakers in the context. In addition, a two-pass decoding strategy is further proposed to fully leverage the contextual modeling ability resulting in a better recognition performance. Experimental results on AliMeeting corpus show that the proposed CASA-ASR model outperforms the original E2E SA-ASR system with a relative improvement of 11.76% in terms of speaker-dependent character error rate.
For speech interaction, voice activity detection (VAD) is often used as a front-end. However, traditional VAD algorithms usually need to wait for a continuous tail silence to reach a preset maximum duration before segmentation, resulting in a large latency that affects user experience. In this paper, we propose a novel semantic VAD for low-latency segmentation. Different from existing methods, a frame-level punctuation prediction task is added to the semantic VAD, and the artificial endpoint is included in the classification category in addition to the often-used speech presence and absence. To enhance the semantic information of the model, we also incorporate an automatic speech recognition (ASR) related semantic loss. Evaluations on an internal dataset show that the proposed method can reduce the average latency by 53.3% without significant deterioration of character error rate in the back-end ASR compared to the traditional VAD approach.