Unleashing the killer corpus: experiences in creating the multi-everything AMI Meeting Corpus

By J. Carletta. Language Resources and Evaluation, 2007

437 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Literature and Literary Theory

View Paper: https://www.semanticscholar.org/paper/0b2da85ac12a9a1e82f1714f2e2dce073e8a158a

How researchers see this paper

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

In Artificial Intelligence

In Literature and Literary Theory

Citations in the corpus

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).”
Pyannote.Audio: Neural Building Blocks for Speaker Diarization (2019, methodology): “…were trained and tested on VoxCeleb [19], all other models (including the full diarization pipeline) were trained, tuned, and tested on three different datasets, covering a wide range of domains: meetings for AMI [20], broadcast news for ETAPE [21], and up to 11 different domains for DIHARD [3].”
Generating SOAP Notes from Doctor-Patient Conversations Using Modular Summarization Techniques (2020, methodology): “Additionally, we evaluate our summarization methods on the AMI dataset (Carletta, 2007), comprised of business meeting transcripts and their summaries.”
Highway long short-term memory RNNS for distant speech recognition (2015, methodology): “We evaluated our models on the AMI meeting corpus [21].”
Encoder-Decoder Based Attractor Calculation for End-to-End Neural Diarization (2021, background): “It can be placed as a downstream task of automatic speech recognition (ASR), in which speaker information is tagged to each transcribed utterance [1]–[3].”
Using Summarization to Discover Argument Facets in Online Idealogical Dialog (2017, background): “Much previous work on dialog summarization focused on extracting phenomena specific to meetings, such as action items or decisions (Murray et al., 2006; Hsueh and Moore, 2008; Whittaker et al., 2012; Janin et al., 2004; Carletta, 2007).”
Recognition of overlapping speech using digital MEMS microphone arrays (2013, methodology): “In those experiments we used circular arrays of a diameter of 20 cm, similar to those used for data collection in the AMI Meetings Corpus [11].”
Interpretation of Multiparty Meetings the AMI and Amida Projects (2008, methodology): “These instrumented meeting rooms were used to record the AMI Meeting Corpus [3], which consists of 100 hours of multimodal meeting recordings, with the different recording streams synchronized to a common timeline.”
Efficient Knowledge Distillation from an Ensemble of Teachers (2017, background): “50% of this training corpora is clean audio from three public corpora - 100 hours from broadcast news, 100 hours from Mixer 6 [15], and 20 hours from the AMI corpus [16] and 30 hours of private speech data.”
Recent trends in deep learning based personality detection (2019, methodology): “Valente et al. (2012) have tested their model on spoken conversations from the AMI corpus dataset (Carletta 2007), which is a collection of meetings captured in specially instrumented meeting rooms along with the audio and video of each of the participants.”
Convolutional Neural Networks for Distant Speech Recognition (2014, background): “The m-th max-pooled band is composed of J related filters pm = [p1,m, . . . , pj,m, . . . , pJ,m] 2 RJ : pj,m = R max r=1 hj,(m 1)N+r , (2) 1The convolution of two vectors of size X and Y may result either in the vector of size X + Y 1 for a full convolution with zero-padding of nonoverlapping…”
End-to-end speaker segmentation for overlap-aware resegmentation (2021, methodology): “OSD AMI [8, 15] DIHARD 3 [9] VoxConverse [11] FA Miss.”
Hierarchical Bayesian Language Models for Conversational Speech Recognition (2010, background): “A second corpus we consider in this paper is a domain-specific meeting corpus—the AMI Meeting Corpus,3 [40] which consists of 100 hours of multimodal meeting recordings with comprehensive annotations at a number of different levels.”
SLURP: A Spoken Language Understanding Resource Package (2020, background): “Multi-ASR is a large-scale system estimated from publicly available acoustic data pooled together – Acoustic data including, among others, LibriSpeech (Panayotov et al., 2015), Switchboard (Godfrey et al., 1992), Fisher (Cieri et al., 2004), CommonVoice (Ardila et al., 2019), AMI (Carletta, 2007) an”
Improved MVDR Beamforming Using Single-Channel Mask Prediction Networks (2016, background): “Recent work on multi-microphone databases such as the AMI database [1, 2, 3] and recent challenges such as REVERB [4], CHiME-2 [5], and CHiME-3 [6] indicate that performing beamforming of multiple microphone data improves performance as compared to other direct techniques for speech recognition.”