The AMI meeting corpus

By Iain McCowan, J. Carletta, Wessel Kraaij, S. Ashby, S. Bourban, Mike Flynn, M. Guillemot, Thomas Hain, J. Kadlec, Vasilis Karaiskos and 7 more. 2005

750 citations recorded by the source corpus.

Subfields: Evaluation & Benchmarks, Speech & Audio

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Evaluation & Benchmarks

In Speech & Audio

Citations in the corpus

SAMSum Corpus: A Human-annotated Dialogue Dataset for Abstractive Summarization (2019, methodology): “The challenges posed by the abstractive dialogue summarization task have been discussed in the literature with regard to AMI meeting corpus (McCowan et al., 2005), e.”
Ego4D: Around the World in 3,000 Hours of Egocentric Video (2021, background): “The TTM task is related to audio-visual speaker detection [7,193] and meeting understanding [21,132,154].”
Gender and Dialect Bias in YouTube’s Automatic Captions (2017, background): “Many other popular speech corpora such as the Numbers corpus (Cole et al., 1995) or the AMI meeting corpus (McCowan et al., 2005) don’t include information on speaker gender or dialect background.”
VoxCeleb: A Large-Scale Speaker Identification Dataset (2017, background): “Datasets obtained from multi-speaker environments include those from recorded meeting data [23, 24], or from audio broadcasts [25].”
CONFIT: Toward Faithful Dialogue Summarization with Linguistically-Informed Contrastive Fine-tuning (2021, methodology): “In recent years, the resurgence of dialogue summarization has attracted significant research attentions (McCowan et al., 2005; Gliwa et al., 2019; Koay et al., 2020; Zhang et al., 2021; Zhong et al., 2021; Zhu et al., 2021; Chen et al., 2021a; Li et al., 2021; Chen et al., 2021c; Fabbri et al., 2021”
Transcribing Meetings With the AMIDA Systems (2012, background): “For a full description the reader is referred to [8] or the documentation available with the corpus.”
Far-Field ASR Without Parallel Data (2016, methodology): “There are three LVCSR tasks [7, 8] designed using the AMI meeting corpora [1].”
FINDINGS OF THE IWSLT 2020 EVALUATION CAMPAIGN (2020, methodology): “The development set contained 2 other ﬁles from Antrecorp, one other ﬁle from the SAO domain and it also included 4 ﬁles from the AMI corpus (Mccowan et al., 2005) to illustrate non-native accents.”
Factorized Hidden Layer Adaptation for Deep Neural Network Based Acoustic Modeling (2016, methodology): “In this paper, we have evaluated the proposed method in three benchmark ASR tasks: Aurora 4 [24], [25], Augmented Multi-party Interaction (AMI) [26], [27] individual headset microphone (IHM) and AMI single distance microphone (SDM).”
Multi-talker ASR for an unknown number of sources: Joint training of source counting, separation and ASR (2020, background): “Overlapping speech is quite common for many scenarios, as the meetings recorded in the AMI corpus [1] show overlap in about 5 % to 10 % of the time, and in informal situations, as recorded in the CHiME-5 [2] challenge data, it can exceed 20 %.”
Ava Active Speaker: An Audio-Visual Dataset for Active Speaker Detection (2019, background): “Meetings data [2, 34] contain spontaneous speech, and non-frontal and occluded faces.”
MediaSum: A Large-scale Media Interview Dataset for Dialogue Summarization (2021, methodology): “Speciﬁcally, we take the pre-trained model UniLM (Dong et al., 2019), ﬁne-tune it on M EDIA S UM , and then train it on datasets for meeting and dialogue summarization: AMI (McCowan et al., 2005), ICSI (Janin et al., 2003) and SAMSum (Gliwa et al., 2019).”
Investigation of transfer learning for ASR using LF-MMI trained neural networks (2017, methodology): “We use several different corpora for our experiments – Switchboard (SWBD), Librispeech [28], WSJ and AMI [29] in both individual headset microphone (IHM) and single distance microphone (SDM) conditions.”
On Multi-Domain Training and Adaptation of End-to-End RNN Acoustic Models for Distant Speech Recognition (2017, methodology): “This study uses the far-field recordings of AMI corpus [16] to examine how environmentally-robust representations are learned by RNN acoustic models.”
Voxceleb: Large-scale speaker verification in the wild (2020, background): “Datasets obtained from multi-speaker environments include those from recorded meeting data (Janin et al., 2003; McCowan et al., 2005), or from audio broadcasts (Bell et al., 2015).”