Advances in automatic meeting record creation and access

Abstract

Oral communication is transient but many important decisions, social contracts and fact 'ndings are 'rst canied out in an oral setup, documented in written form and later retrieved. At Carnegie Mellons University s Interactive Systems Laboratories we have been experimenting with the documentation of meetings. T h s paper summarizes part of the progress that we have made in this test bed, speci'cally on the question of automatic transcription using LVCSR, information access using non-keyword based methods, summarization and user interfaces. The system is capable to automatically construct a searchable and browsable audiovisual database of meetings and provide access to these records.

Toward Joint Segmentation and Classification of Dialog Acts in Multiparty Meetings

“…To support higher-level tasks such as information retrieval and summarization [1,2], the input speech signal must be segmented into meaningful units, for example dialog acts (DAs). The five DA types used in this work are statements, questions, backchannels, floorgrabbers, and disruptions.…”

Navigating Multimodal Meeting Recordings with the Meeting Miner

“…Existing systems include the Meeting Browser [5] in which automatic speech recognition (ASR) transcripts are used for meeting navigation and summarization. The SCAN (Spoken Content based Audio Navigation) system [6] uses acoustic and prosodic features for audio segmentation and a number of information retrieval techniques applied on ASR transcripts for speech recording indexing.…”

“…Thus, the idea behind using extended overlap as a basis for information retrieval is based on our observations that in accordance with the natural structure of discourse, media intervals in close time proximity are often semantically related. This can be viewed as the questionanswer pair paradigm [5] where adjacent speech exchanges are more informative jointly than when considered on their own. A more detailed discussion on media stream interactions and implications for media intervals temporal segmentation can be found in [12].…”

Exploring the Structure of Media Stream Interactions for Multimedia Browsing

“…Such linking patterns can be represented by tree structures, as illustrated in Figure 2. This kind of navigation structure can evolve along several dimensions, beyond what is supported by meeting browsers based on a linear access metaphor, including those supported by sophisticated modality translation components [16,17].…”