Thai Binh Nguyen

Christian Huber, Tu Anh Dinh, Carlos Mullov et al.

The challenge of low-latency speech translation has recently draw significant interest in the research community as shown by several publications and shared tasks. Therefore, it is essential to evaluate these different approaches in realistic scenarios. However, currently only specific aspects of the systems are evaluated and often it is not possible to compare different approaches. In this work, we propose the first framework to perform and evaluate the various aspects of low-latency speech translation under realistic conditions. The evaluation is carried out in an end-to-end fashion. This includes the segmentation of the audio as well as the run-time of the different components. Secondly, we compare different approaches to low-latency speech translation using this framework. We evaluate models with the option to revise the output as well as methods with fixed output. Furthermore, we directly compare state-of-the-art cascaded as well as end-to-end systems. Finally, the framework allows to automatically evaluate the translation quality as well as latency and also provides a web interface to show the low-latency model outputs to the user.

Danni Liu, Thai Binh Nguyen, Sai Koneru et al.

Many existing speech translation benchmarks focus on native-English speech in high-quality recording conditions, which often do not match the conditions in real-life use-cases. In this paper, we describe our speech translation system for the multilingual track of IWSLT 2023, which evaluates translation quality on scientific conference talks. The test condition features accented input speech and terminology-dense contents. The task requires translation into 10 languages of varying amounts of resources. In absence of training data from the target domain, we use a retrieval-based approach (kNN-MT) for effective adaptation (+0.8 BLEU for speech translation). We also use adapters to easily integrate incremental training data from data augmentation, and show that it matches the performance of re-training. We observe that cascaded systems are more easily adaptable towards specific target domains, due to their separate modules. Our cascaded speech system substantially outperforms its end-to-end counterpart on scientific talk translation, although their performance remains similar on TED talks.

Thai-Binh Nguyen, Alexander Waibel

This paper presents an end-to-end model designed to improve automatic speech recognition (ASR) for a particular speaker in a crowded, noisy environment. The model utilizes a single-channel speech enhancement module that isolates the speaker's voice from background noise (ConVoiFilter) and an ASR module. The model can decrease ASR's word error rate (WER) from 80% to 26.4% through this approach. Typically, these two components are adjusted independently due to variations in data requirements. However, speech enhancement can create anomalies that decrease ASR efficiency. By implementing a joint fine-tuning strategy, the model can reduce the WER from 26.4% in separate tuning to 14.5% in joint tuning. We openly share our pre-trained model to foster further research hf.co/nguyenvulebinh/voice-filter.

Thi-Hai-Yen Vuong, Hai-Long Nguyen, Tan-Minh Nguyen et al.

This paper presents the NOWJ team's approach to the COLIEE 2023 Competition, which focuses on advancing legal information processing techniques and applying them to real-world legal scenarios. Our team tackles the four tasks in the competition, which involve legal case retrieval, legal case entailment, statute law retrieval, and legal textual entailment. We employ state-of-the-art machine learning models and innovative approaches, such as BERT, Longformer, BM25-ranking algorithm, and multi-task learning models. Although our team did not achieve state-of-the-art results, our findings provide valuable insights and pave the way for future improvements in legal information processing.

Tan-Minh Nguyen, Thai-Binh Nguyen, Hoang-Trung Nguyen et al.

Multi-document summarization is challenging because the summaries should not only describe the most important information from all documents but also provide a coherent interpretation of the documents. This paper proposes a method for multi-document summarization based on cluster similarity. In the extractive method we use hybrid model based on a modified version of the PageRank algorithm and a text correlation considerations mechanism. After generating summaries by selecting the most important sentences from each cluster, we apply BARTpho and ViT5 to construct the abstractive models. Both extractive and abstractive approaches were considered in this study. The proposed method achieves competitive results in VLSP 2022 competition.

Fabian Retkowski, Maike Züfle, Thai Binh Nguyen et al.

Audio chaptering, the task of automatically segmenting long-form audio into coherent sections, is increasingly important for navigating podcasts, lectures, and videos. Despite its relevance, research remains limited and text-based, leaving key questions unresolved about leveraging audio information, handling ASR errors, and transcript-free evaluation. We address these gaps through three contributions: (1) a systematic comparison between text-based models with acoustic features, a novel audio-only architecture (AudioSeg) operating on learned audio representations, and multimodal LLMs; (2) empirical analysis of factors affecting performance, including transcript quality, acoustic features, duration, and speaker composition; and (3) formalized evaluation protocols contrasting transcript-dependent text-space protocols with transcript-invariant time-space protocols. Our experiments on YTSeg reveal that AudioSeg substantially outperforms text-based approaches, pauses provide the largest acoustic gains, and MLLMs remain limited by context length and weak instruction following, yet MLLMs are promising on shorter audio.

Thai-Binh Nguyen, Alexander Waibel

Speaker-attributed automatic speech recognition (SA-ASR) aims to transcribe speech while assigning transcripts to the corresponding speakers accurately. Existing methods often rely on complex modular systems or require extensive fine-tuning of joint modules, limiting their adaptability and general efficiency. This paper introduces a novel approach, leveraging a frozen multilingual ASR model to incorporate speaker attribution into the transcriptions, using only standard monolingual ASR datasets. Our method involves training a speaker module to predict speaker embeddings based on weak labels without requiring additional ASR model modifications. Despite being trained exclusively with non-overlapping monolingual data, our approach effectively extracts speaker attributes across diverse multilingual datasets, including those with overlapping speech. Experimental results demonstrate competitive performance compared to strong baselines, highlighting the model's robustness and potential for practical applications.

Thai-Binh Nguyen, Ngoc-Quan Pham, Alexander Waibel

Audio-Visual Speech Recognition (AVSR) offers a robust solution for speech recognition in challenging environments, such as cocktail-party scenarios, where relying solely on audio proves insufficient. However, current AVSR models are often optimized for idealized scenarios with consistently active speakers, overlooking the complexities of real-world settings that include both speaking and silent facial segments. This study addresses this gap by introducing a novel audio-visual cocktail-party dataset designed to benchmark current AVSR systems and highlight the limitations of prior approaches in realistic noisy conditions. Additionally, we contribute a 1526-hour AVSR dataset comprising both talking-face and silent-face segments, enabling significant performance gains in cocktail-party environments. Our approach reduces WER by 67% relative to the state-of-the-art, reducing WER from 119% to 39.2% in extreme noise, without relying on explicit segmentation cues.

Sai Koneru, Maike Züfle, Thai-Binh Nguyen et al.

The scope of the International Workshop on Spoken Language Translation (IWSLT) has recently broadened beyond traditional Speech Translation (ST) to encompass a wider array of tasks, including Speech Question Answering and Summarization. This shift is partly driven by the growing capabilities of modern systems, particularly with the success of Large Language Models (LLMs). In this paper, we present the Karlsruhe Institute of Technology's submissions for the Offline ST and Instruction Following (IF) tracks, where we leverage LLMs to enhance performance across all tasks. For the Offline ST track, we propose a pipeline that employs multiple automatic speech recognition systems, whose outputs are fused using an LLM with document-level context. This is followed by a two-step translation process, incorporating additional refinement step to improve translation quality. For the IF track, we develop an end-to-end model that integrates a speech encoder with an LLM to perform a wide range of instruction-following tasks. We complement it with a final document-level refinement stage to further enhance output quality by using contextual information.

Thai-Binh Nguyen, Thi Van Nguyen, Quoc Truong Do et al.

Audio-Visual Speech Recognition (AVSR) has gained significant attention recently due to its robustness against noise, which often challenges conventional speech recognition systems that rely solely on audio features. Despite this advantage, AVSR models remain limited by the scarcity of extensive datasets, especially for most languages beyond English. Automated data collection offers a promising solution. This work presents a practical approach to generate AVSR datasets from raw video, refining existing techniques for improved efficiency and accessibility. We demonstrate its broad applicability by developing a baseline AVSR model for Vietnamese. Experiments show the automatically collected dataset enables a strong baseline, achieving competitive performance with robust ASR in clean conditions and significantly outperforming them in noisy environments like cocktail parties. This efficient method provides a pathway to expand AVSR to more languages, particularly under-resourced ones.

Fevziye Irem Eyiokur, Christian Huber, Thai-Binh Nguyen et al.

In this paper, we report on communication experiments conducted in the summer of 2022 during a deep dive to the wreck of the Titanic. Radio transmission is not possible in deep sea water, and communication links rely on sonar signals. Due to the low bandwidth of sonar signals and the need to communicate readable data, text messaging is used in deep-sea missions. In this paper, we report results and experiences from a messaging system that converts speech to text in a submarine, sends text messages to the surface, and reconstructs those messages as synthetic lip-synchronous videos of the speakers. The resulting system was tested during an actual dive to Titanic in the summer of 2022. We achieved an acceptable latency for a system of such complexity as well as good quality. The system demonstration video can be found at the following link: https://youtu.be/C4lyM86-5Ig

Sai Koneru, Thai-Binh Nguyen, Ngoc-Quan Pham et al.

Large Language Models (LLMs) are currently under exploration for various tasks, including Automatic Speech Recognition (ASR), Machine Translation (MT), and even End-to-End Speech Translation (ST). In this paper, we present KIT's offline submission in the constrained + LLM track by incorporating recently proposed techniques that can be added to any cascaded speech translation. Specifically, we integrate Mistral-7B\footnote{mistralai/Mistral-7B-Instruct-v0.1} into our system to enhance it in two ways. Firstly, we refine the ASR outputs by utilizing the N-best lists generated by our system and fine-tuning the LLM to predict the transcript accurately. Secondly, we refine the MT outputs at the document level by fine-tuning the LLM, leveraging both ASR and MT predictions to improve translation quality. We find that integrating the LLM into the ASR and MT systems results in an absolute improvement of $0.3\%$ in Word Error Rate and $0.65\%$ in COMET for tst2019 test set. In challenging test sets with overlapping speakers and background noise, we find that integrating LLM is not beneficial due to poor ASR performance. Here, we use ASR with chunked long-form decoding to improve context usage that may be unavailable when transcribing with Voice Activity Detection segmentation alone.

Thai Binh Nguyen, Quang Minh Nguyen, Thi Thu Hien Nguyen et al.

Studies on the Named Entity Recognition (NER) task have shown outstanding results that reach human parity on input texts with correct text formattings, such as with proper punctuation and capitalization. However, such conditions are not available in applications where the input is speech, because the text is generated from a speech recognition system (ASR), and that the system does not consider the text formatting. In this paper, we (1) presented the first Vietnamese speech dataset for NER task, and (2) the first pre-trained public large-scale monolingual language model for Vietnamese that achieved the new state-of-the-art for the Vietnamese NER task by 1.3% absolute F1 score comparing to the latest study. And finally, (3) we proposed a new pipeline for NER task from speech that overcomes the text formatting problem by introducing a text capitalization and punctuation recovery model (CaPu) into the pipeline. The model takes input text from an ASR system and performs two tasks at the same time, producing proper text formatting that helps to improve NER performance. Experimental results indicated that the CaPu model helps to improve by nearly 4% of F1-score.

Thai Binh Nguyen, Quang Minh Nguyen, Thu Hien Nguyen et al.

Nowadays, Social network sites (SNSs) such as Facebook, Twitter are common places where people show their opinions, sentiments and share information with others. However, some people use SNSs to post abuse and harassment threats in order to prevent other SNSs users from expressing themselves as well as seeking different opinions. To deal with this problem, SNSs have to use a lot of resources including people to clean the aforementioned content. In this paper, we propose a supervised learning model based on the ensemble method to solve the problem of detecting hate content on SNSs in order to make conversations on SNSs more effective. Our proposed model got the first place for public dashboard with 0.730 F1 macro-score and the third place with 0.584 F1 macro-score for private dashboard at the sixth international workshop on Vietnamese Language and Speech Processing 2019.

Quang Minh Nguyen, Thai Binh Nguyen, Ngoc Phuong Pham et al.

Automatic Speech Recognition (ASR) systems have been evolving quickly and reaching human parity in certain cases. The systems usually perform pretty well on reading style and clean speech, however, most of the available systems suffer from situation where the speaking style is conversation and in noisy environments. It is not straight-forward to tackle such problems due to difficulties in data collection for both speech and text. In this paper, we attempt to mitigate the problems using language models combination techniques that allows us to utilize both large amount of writing style text and small number of conversation text data. Evaluation on the VLSP 2019 ASR challenges showed that our system achieved 4.85% WER on the VLSP 2018 and 15.09% WER on the VLSP 2019 data sets.

ThaiBinh Nguyen, Atsuhiro Takasu

Matrix factorization (MF) is one of the most efficient methods for rating predictions. MF learns user and item representations by factorizing the user-item rating matrix. Further, textual contents are integrated to conventional MF to address the cold-start problem. However, the textual contents do not reflect all aspects of the items. In this paper, we propose a model that leverages the information hidden in the item co-click (i.e., items that are often clicked together by a user) into learning item representations. We develop TCMF (Textual Co Matrix Factorization) that learns the user and item representations jointly from the user-item matrix, textual contents and item co-click matrix built from click data. Item co-click information captures the relationships between items which are not captured via textual contents. The experiments on two real-world datasets MovieTweetings, and Bookcrossing) demonstrate that our method outperforms competing methods in terms of rating prediction. Further, we show that the proposed model can learn effective item representations by comparing with state-of-the-art methods in classification task which uses the item representations as input vectors.

ThaiBinh Nguyen, Atsuhiro Takasu

Matrix factorization is one of the most efficient approaches in recommender systems. However, such algorithms, which rely on the interactions between users and items, perform poorly for "cold-users" (users with little history of such interactions) and at capturing the relationships between closely related items. To address these problems, we propose a neural personalized embedding (NPE) model, which improves the recommendation performance for cold-users and can learn effective representations of items. It models a user's click to an item in two terms: the personal preference of the user for the item, and the relationships between this item and other items clicked by the user. We show that NPE outperforms competing methods for top-N recommendations, specially for cold-user recommendations. We also performed a qualitative analysis that shows the effectiveness of the representations learned by the model.

ThaiBinh Nguyen, Kenro Aihara, Atsuhiro Takasu

Collaborative filtering is the most popular approach for recommender systems. One way to perform collaborative filtering is matrix factorization, which characterizes user preferences and item attributes using latent vectors. These latent vectors are good at capturing global features of users and items but are not strong in capturing local relationships between users or between items. In this work, we propose a method to extract the relationships between items and embed them into the latent vectors of the factorization model. This combines two worlds: matrix factorization for collaborative filtering and item embed- ding, a similar concept to word embedding in language processing. Our experiments on three real-world datasets show that our proposed method outperforms competing methods on top-n recommendation tasks.

ThaiBinh Nguyen, Atsuhiro Takasu

One of the most efficient methods in collaborative filtering is matrix factorization, which finds the latent vector representations of users and items based on the ratings of users to items. However, a matrix factorization based algorithm suffers from the cold-start problem: it cannot find latent vectors for items to which previous ratings are not available. This paper utilizes click data, which can be collected in abundance, to address the cold-start problem. We propose a probabilistic item embedding model that learns item representations from click data, and a model named EMB-MF, that connects it with a probabilistic matrix factorization for rating prediction. The experiments on three real-world datasets demonstrate that the proposed model is not only effective in recommending items with no previous ratings, but also outperforms competing methods, especially when the data is very sparse.