arXiv subjects

Heyang Liu

Publications and source records attributed to Heyang Liu.

23 records · Page 2Linked to original sources

arXiv2024

M$^3$AV: A Multimodal, Multigenre, and Multipurpose Audio-Visual Academic Lecture Dataset

Publishing open-source academic video recordings is an emergent and prevalent approach to sharing knowledge online. Such videos carry rich multimodal information including speech, the facial and body movements of the speakers, as well as the texts and pictures in the slides and possibly even the papers. Although multiple academic video datasets have been constructed and released, few of them support both multimodal content recognition and understanding tasks, which is partially due to the lack of high-quality human annotations. In this paper, we propose a novel multimodal, multigenre, and multipurpose audio-visual academic lecture dataset (M$^3$AV), which has almost 367 hours of videos from five sources covering computer science, mathematics, and medical and biology topics. With high-quality human annotations of the slide text and spoken words, in particular high-valued name entities, the dataset can be used for multiple audio-visual recognition and understanding tasks. Evaluations performed on contextual speech recognition, speech synthesis, and slide and script generation tasks demonstrate that the diversity of M$^3$AV makes it a challenging dataset.

cs.CL↗

arXiv2024

MM-SAP: A Comprehensive Benchmark for Assessing Self-Awareness of Multimodal Large Language Models in Perception

Recent advancements in Multimodal Large Language Models (MLLMs) have demonstrated exceptional capabilities in visual perception and understanding. However, these models also suffer from hallucinations, which limit their reliability as AI systems. We believe that these hallucinations are partially due to the models' struggle with understanding what they can and cannot perceive from images, a capability we refer to as self-awareness in perception. Despite its importance, this aspect of MLLMs has been overlooked in prior studies. In this paper, we aim to define and evaluate the self-awareness of MLLMs in perception. To do this, we first introduce the knowledge quadrant in perception, which helps define what MLLMs know and do not know about images. Using this framework, we propose a novel benchmark, the Self-Awareness in Perception for MLLMs (MM-SAP), specifically designed to assess this capability. We apply MM-SAP to a variety of popular MLLMs, offering a comprehensive analysis of their self-awareness and providing detailed insights. The experiment results reveal that current MLLMs possess limited self-awareness capabilities, pointing to a crucial area for future advancement in the development of trustworthy MLLMs. Code and data are available at https://github.com/YHWmz/MM-SAP.

cs.CV↗

arXiv2024

LibriSQA: A Novel Dataset and Framework for Spoken Question Answering with Large Language Models

While Large Language Models (LLMs) have demonstrated commendable performance across a myriad of domains and tasks, existing LLMs still exhibit a palpable deficit in handling multimodal functionalities, especially for the Spoken Question Answering (SQA) task which necessitates precise alignment and deep interaction between speech and text features. To address the SQA challenge on LLMs, we initially curated the free-form and open-ended LibriSQA dataset from Librispeech, comprising Part I with natural conversational formats and Part II encompassing multiple-choice questions followed by answers and analytical segments. Both parts collectively include 107k SQA pairs that cover various topics. Given the evident paucity of existing speech-text LLMs, we propose a lightweight, end-to-end framework to execute the SQA task on the LibriSQA, witnessing significant results. By reforming ASR into the SQA format, we further substantiate our framework's capability in handling ASR tasks. Our empirical findings bolster the LLMs' aptitude for aligning and comprehending multimodal information, paving the way for the development of universal multimodal LLMs. The dataset and demo can be found at https://github.com/ZihanZhaoSJTU/LibriSQA.

cs.CL↗

arXiv2024

Post-decoder Biasing for End-to-End Speech Recognition of Multi-turn Medical Interview

End-to-end (E2E) approach is gradually replacing hybrid models for automatic speech recognition (ASR) tasks. However, the optimization of E2E models lacks an intuitive method for handling decoding shifts, especially in scenarios with a large number of domain-specific rare words that hold specific important meanings. Furthermore, the absence of knowledge-intensive speech datasets in academia has been a significant limiting factor, and the commonly used speech corpora exhibit significant disparities with realistic conversation. To address these challenges, we present Medical Interview (MED-IT), a multi-turn consultation speech dataset that contains a substantial number of knowledge-intensive named entities. We also explore methods to enhance the recognition performance of rare words for E2E models. We propose a novel approach, post-decoder biasing, which constructs a transform probability matrix based on the distribution of training transcriptions. This guides the model to prioritize recognizing words in the biasing list. In our experiments, for subsets of rare words appearing in the training speech between 10 and 20 times, and between 1 and 5 times, the proposed method achieves a relative improvement of 9.3% and 5.1%, respectively.

cs.CL↗

arXiv2022

Multi-wavelength Campaign on the Super-Eddington NLS1 RX J0134.2-4258 -- I. Peculiar X-ray Spectra and Variability

We have conducted a new long-term multi-wavelength campaign on one of the most super- Eddington narrow-line Seyfert 1s (NLS1s) known, namely RX J0134.2-4258. In this first paper, we report deep simultaneous X-ray observations performed by XMM-Newton and NuSTAR in 2019-12-19, during which RX J0134.2-4258 was fortuitously at one of its lowest X-ray flux states. However, there is a clear rise above 4 keV which implies that the intrinsic source flux may be higher. The X-ray spectra observed between 1996 and 2019 show drastic variability, probably due to complex, variable absorption along the line of sight. Unusually, the soft X-ray excess appears extremely weak in all these spectra, even when the hard X-ray spectrum has a steep slope of $Γ\simeq 2.2$. We explore the spectral-timing properties of the new (low X-ray flux) and archival (high X-ray flux) XMM-Newton data, fitting their time-average, rms and lag spectra simultaneously. The variability spectra indicate the presence of a very weak soft X-ray Comptonisation component, whose shape is similar to the soft excess in normal super-Eddington NLS1s, but with flux relative to the power law which is lower by more than one order of magnitude. Above 4 keV the low-flux data are dominated by a different component, which lags with respect to the lower energy emission. This is consistent with an origin of reflection or partial covering absorption from low ionization material located within 100 Rg. We interpret this as a further indication of the presence of a clumpy disc wind.

astro-ph.HE↗