Academia.edu no longer supports Internet Explorer.
To browse Academia.edu and the wider internet faster and more securely, please take a few seconds to upgrade your browser.
2005
...
13 pages
Abstract The disclosure of audio-visual meeting recordings is a new challenging domain studied by several large scale research projects in Europe and the US. Automatic meeting summarization is one of the functionalities studied. In this paper we report the results of a feasibility study on a subtask, namely the summarization of meeting transcripts.
AI
Meetings have lower information density and lack clear structural cues, making summarization difficult. Unlike news articles, conversations often include many disfluencies and trivial backchannels that complicate extraction.
The Maximum Entropy model achieved a relatively low F-measure of 0.505, indicating substantial room for improvement. Despite the performance limitations, it enhanced baseline results by approximately 20%, showcasing its effectiveness.
Analysis revealed that frequent speakers often contribute more significant information, underscoring the relevance of speaker dynamics. Features indicating important speakers improved model performance by capturing vital segments.
The training dataset consisted of six manually annotated meetings from the ICSI corpus, taking 12-14 hours to annotate each. Approximately 22,000 segments were evaluated using a ternary importance scale for summarization.
Future work may involve incorporating lexical chains to enhance topic detection and context relevance. This could lead to better summaries by connecting critical sentences based on topical coherence.
Research Challenges in Information Science, 2020
A lot of research has been conducted all over the world in the domain of automatic text summarization and more specifically using machine learning techniques. Many state of the art prototypes partially solve this problem so we decided to use some of them to build a tool for automatic generation of meeting minutes. In fact, this was not an easy work and this paper presents various experiments that we did using Deep Learning, GANs and Transformers to achieve this goal as well as dead ends we have encountered during this study. We think providing such a feedback may be useful to other researchers who would like to undertake the same type of work to allow them to know where to go and where not to go.
International Journal of Scientific Research in Science, Engineering and Technology, 2022
This paper describes the process for automatic generation of minutes of meetings using Machine Learning algorithms and Natural Language Processing techniques. Minutes of meetings are a record which are used to keep official summaries of all meetings conducted within a company or organization. Automatic generation of minutes of meeting is a challenging issue and has gathered a huge amount of interest over the last few years due to its applications. Initially, we study previous research papers to understand existing techniques used for the purpose. Techniques such as AMBOC Model, BART Summarizer, HMNet Model, MSCG are employed for detecting useful and informative action items from audio files. Then we explore Machine Learning models such as SVM, HMM which are clubbed with the majority of methods for classification and summarization of the words given by above mentioned models to generate an informative summary for the user.
2014
In this paper, we present an automatic abstractive summarization system of meeting conversations. Our system extends a novel multi-sentence fusion algorithm in order to generate abstract templates. It also leverages the relationship between summaries and their source meeting transcripts to select the best templates for generating abstractive summaries of meetings. Our manual and automatic evaluation results demonstrate the success of our system in achieving higher scores both in readability and informativeness.
Computer Speech & Language, 2010
Meeting summarization provides a concise and informative summary for the lengthy meetings and is an effective tool for efficient information access. In this paper, we focus on extractive summarization, where salient sentences are selected from the meeting transcripts to form a summary. We adopt a supervised learning approach for this task and use a classifier to determine whether to select a sentence in the summary based on a rich set of features. We address two important problems associated with this supervised classification approach. First we propose different sampling methods to deal with the imbalanced data problem for this task where the summary sentences are the minority class. Second, in order to account for human disagreement for summary annotation, we reframe the extractive summarization task using a regression scheme instead of binary classification. We evaluate our approaches using the ICSI meeting corpus on both the human transcripts and speech recognition output, and show performance improvement using different sampling methods and regression model.
2009 IEEE Workshop on Automatic Speech Recognition & Understanding, 2009
Speech contains additional information than text that can be valuable for automatic speech summarization. In this paper, we evaluate how to effectively use acoustic/prosodic features for extractive meeting summarization, and how to integrate prosodic features with lexical and structural information for further improvement. To properly represent prosodic features, we propose different normalization methods based on speaker, topic, or local context information. Our experimental results show that using only the prosodic features we achieve better performance than using the non-prosodic information on both the human transcripts and recognition output. In addition, a decision-level combination of the prosodic and non-prosodic features yields further gain, outperforming the individual models.
ACM SIGIR Forum, 2021
The SummDial special session on summarization of dialogues and multi-party meetings was held virtually within the SIGDial 2021 conference on July 29, 2021. SummDial @ SIGDial 2021 aimed to bring together the speech, dialogue, and summarization communities to foster cross-pollination of ideas and fuel the discussions/collaborations to attempt this crucial and timely problem. When the pandemic has restricted most of our in-person interactions, the current scenario has forced people to go virtual, resulting in an information overload from frequent dialogues and meetings in the virtual environment. Summarization could help reduce the cognitive burden on the participants; however, multi-party speech summarization comes with its own set of challenges. The SummDial special session aimed to leverage the community intelligence to find effective solutions while also brainstorming the future of AI interventions in meetings and dialogues. We report the findings of the special session in this ar...
International Journal of Computer Applications
Automated meeting minutes are a technological solution that can revolutionize the way organizations conduct meetings. Traditional meetings can be time-consuming, and inefficient, and often result in incomplete or inaccurate meeting minutes. Automated meeting minutes can address these issues by providing several advantages over traditional methods. The purpose of this research paper is to explore the development of an automated system for generating meeting minutes in the IT industry. The current manual process of creating meeting minutes can be time-consuming and prone to errors. The proposed system will use Speech Emotion Recognition (SER) technology to identify the emotional levels of meeting attendees, such as anger, joy, or neutral emotion, and analyze meeting progress using those emotional levels. The system will generate a meeting summary report that includes meeting objectives, attendance, decisions, issues, information on the next meeting, action items, and progress reports. It also aims to address the challenges faced by organizations when manually creating meeting minutes by automating the process. The proposed system will include important components such as meeting objectives, attendance, decisions, and issues, information on the next meeting, action items, progress reports, and emotional sensitivity analysis. The purpose of this research is to provide users in the IT industry with an interactive analysis of the meeting summary from various standpoints, including generating a meeting minute according to the agenda, identifying the action items of each attendee, tracking the progress of previous actions, and analyzing emotional sensitivity. The proposed system will utilize Speech Emotion Recognition (SER) technology to identify the emotional levels of meeting attendees, such as anger, joy, or neutral emotion, and analyze meeting progress using those emotional levels. The research aims to improve the efficiency and accuracy of meeting minutes in the IT industry and provide a more effective way of conducting meetings.
2017 IEEE 21st International Conference on Computer Supported Cooperative Work in Design (CSCWD), 2017
Organisations now frequently rely on virtual collaboration through the use of computer technology. After a sequence of meetings, participants may only need to refer to the most important points rather than the whole meeting proceedings. This paper addresses the need for automated meeting summarisation in virtual meeting systems. An extraction approach to summarisation is adopted and a new algorithm is proposed by extending the TextRank algorithm to include constructs representing the structure of the meeting. This helps extract the most relevant sentences from the meeting transcript. The proposed method was evaluated in the context of student-tutor meetings. Results show that harnessing and utilising the structure of a virtual meeting can lead to more relevant automated summaries.
arXiv (Cornell University), 2023
This paper studies how to effectively build meeting summarization systems for real-world usage using large language models (LLMs). For this purpose, we conduct an extensive evaluation and comparison of various closed-source and open-source LLMs, namely, GPT-4, GPT-3.5, PaLM-2, and LLaMA-2. Our findings reveal that most closed-source LLMs are generally better in terms of performance. However, much smaller open-source models like LLaMA-2 (7B and 13B) could still achieve performance comparable to the large closed-source models even in zero-shot scenarios. Considering the privacy concerns of closed-source models for only being accessible via API, alongside the high cost associated with using fine-tuned versions of the closed-source models, the opensource models that can achieve competitive performance are more advantageous for industrial use. Balancing performance with associated costs and privacy concerns, the LLaMA-2-7B model looks more promising for industrial usage. In sum, this paper offers practical insights on using LLMs for real-world business meeting summarization, shedding light on the trade-offs between performance and cost.
INTERSPEECH, 2022
In this paper, we propose a novel architecture for direct extractive speech-to-speech summarization, ESSumm, which is an unsupervised model without dependence on intermediate transcribed text. Different from previous methods with text presentation, we are aimed at generating a summary directly from speech without transcription. First, a set of smaller speech segments are extracted based on speech signal's acoustic features. For each candidate speech segment, a distance-based summarization confidence score is designed for latent speech representation measure. Specifically, we leverage the off-the-shelf self-supervised convolutional neural network to extract the deep speech features from raw audio. Our approach automatically predicts the optimal sequence of speech segments that capture the key information with a target summary length. Extensive results on two well-known meeting datasets (AMI and ICSI corpora) show the effectiveness of our direct speech-based method to improve the summarization quality with untranscribed data. We also observe that our unsupervised speech-based method even performs on par with recent transcript-based summarization approaches, where extra speech recognition is required.
Loading Preview
Sorry, preview is currently unavailable. You can download the paper by clicking the button above.
Proceedings of the Conference on Empirical Methods in Natural Language Processing - EMNLP '08, 2008
In this paper we describe research on summarizing conversations in the meetings and emails domains. We introduce a conversation summarization system that works in multiple domains utilizing general conversational features, and compare our results with domain-dependent systems for meeting and email data. We find that by treating meetings and emails as conversations with general conversational features in common, we can achieve competitive results with state-of-theart systems that rely on more domain-specific features.
2008
Abstract Significant research efforts have been devoted to speech summarization, including automatic approaches and evaluation metrics. However, a fundamental problem about what summaries are for the speech data and whether humans agree with each other remains unclear. This paper performs an analysis of human annotated extractive summaries using the ICSI meeting corpus with an aim to examine their consistency and the factors impacting human agreement.
arXiv (Cornell University), 2023
As the number of recorded meetings increases, it becomes increasingly important to utilize summarization technology to create useful summaries of these recordings. However, there is a crucial lack of annotated meeting corpora for developing this technology, as it can be hard to collect meetings, especially when the topics discussed are confidential. Furthermore, meeting summaries written by experienced writers are scarce, making it hard for abstractive summarizers to produce sensible output without a reliable reference. This lack of annotated corpora has hindered the development of meeting summarization technology. In this paper, we present MeetingBank, a new benchmark dataset of city council meetings over the past decade. Meeting-Bank is unique among other meeting corpora due to its divide-and-conquer approach, which involves dividing professionally written meeting minutes into shorter passages and aligning them with specific segments of the meeting. This breaks down the process of summarizing a lengthy meeting into smaller, more manageable tasks. The dataset provides a new testbed of various meeting summarization systems and also allows the public to gain insight into how council decisions are made. We make the collection, including meeting video links, transcripts, reference summaries, agenda, and other metadata, publicly available to facilitate the development of better meeting summarization techniques. 1
Advances in Intelligent Systems and Computing, 2020
The appearance of various platforms such as YouTube, Dailymotion and Google Video has a major role in the increasing of the number of videos available on the Internet. For example, more than 15000 video sequences are seen every day on Dailymotion. Consequently, the huge gathered amount of data constitutes a big scientific challenge for managing the underlying knowledge. Particularly, data summarization aims to extract concise abstracts from different types of documents. In the context of this paper, we are interested in summarizing meetings' data. As the quality of video analyzing's output highly depends on the type of data, we propose to establish our own framework for this end. The main goal of our study is to use textual data extracted from Automatic Speech Recognition (ASR) transcriptions of the AMI corpus to give a fully unsupervised summarized version of meeting sequences. Our contribution, called Weighted Histogram for ASR Transcriptions (WHASRT), adopts an extractive, free of annotations and dictionary-based approach. An exhaustive comparative study demonstrates that our method ensured competitive results with the ranking-based methods. The experimental results showed an enhanced performance over the existing clustering-based methods.
Proceedings of the DARPA ..., 1998
2009 IEEE International Conference on Acoustics, Speech and Signal Processing, 2009
We introduce a model for extractive meeting summarization based on the hypothesis that utterances convey bits of information, or concepts. Using keyphrases as concepts weighted by frequency, and an integer linear program to determine the best set of utterances, that is, covering as many concepts as possible while satisfying a length constraint, we achieve ROUGE scores at least as good as a ROUGEbased oracle derived from human summaries. This brings us to a critical discussion of ROUGE and the future of extractive meeting summarization.
Proceedings of the 46th Annual Meeting of the ..., 2008
Automatic summarization evaluation is critical to the development of summarization systems. While ROUGE has been shown to correlate well with human evaluation for content match in text summarization, there are many characteristics in multiparty meeting domain, which may pose potential problems to ROUGE. In this paper, we carefully examine how well the ROUGE scores correlate with human evaluation for extractive meeting summarization. Our experiments show that generally the correlation is rather low, but a significantly better correlation can be obtained by accounting for several unique meeting characteristics, such as disfluencies and speaker information, especially when evaluating system-generated summaries.
2010
In this paper we present a complete system for automatically generating natural language abstracts of meeting conversations. This system is comprised of components relating to interpretation of the meeting documents according to a meeting ontology, transformation or content selection from that source representation to a summary representation, and generation of new summary text. In a formative user study, we compare this approach to gold-standard human abstracts and extracts to gauge the usefulness of the different summary types for browsing meeting conversations. We find that our automatically generated summaries are ranked significantly higher than human-selected extracts on coherence and usability criteria. More generally, users demonstrate a strong preference for abstract-style summaries over extracts.
International Journal of Computer Science and Mobile Computing (IJCSMC), 2023
The majority of public works are carried out online as a result of the coronavirus disease (COVID19) pandemic. Many job interviews, primary health care consultations, and company meetings are conducted entirely online, and universities all over the world have switched to online teaching. Applications for online meetings like Microsoft Team and Google Meet are widely available on the market. We approach a topic with numerous practical applications in this work. A method that takes recorded video as input and produces identical written or audio summaries as output is the subject of this paper. Additionally, the proposed method can be used to create meeting minutes, subtitle or scribe entertainment videos, and create lecture notes from lecture videos. The audio track of a video is converted into text by the proposed system. We also used a text summarization algorithm to create text summaries. Users of the system have the option of using a text summary or making an audio output that matches the text summary. The proposed strategy is carried out in Python and the proposed conspire is assessed utilizing brief recordings got from YouTube. The proposed method is manually validated on a set of uploaded videos because there are neither specific datasets or benchmarks available to evaluate efficacy nor any benchmarks to use in related studies.