Main Page

From Speechwiki
Jump to: navigation, search

Welcome to the Speech Group Wiki page. These pages are collaborative. That means that any user can modify the content and even add new pages. It is very easy to do and you don't need to know any HTML. For info about how to edit Wiki pages see Help:Editing. To just play around, try the Sandbox.

General Info


Technical Info


  • LLNL Deep and Wide -- Deep and wide neural nets for ASR with Lawrence Livermore National Labs.
  • Swordfish -- Babel Swordfish project.
  • RATS -- Robust Automatic Transcription of Speech
  • OUCH -- Outing Unfortunate Characteristics of HMMs
  • Multi -- Progress and other comments on Parrallel Multimodal and Acoustic Processing of "Wild" Data
  • AFRL - Open set speaker recognition in noisy conditions.
  • Microsoft - Microsoft project: Prosody toolkit and applications
  • SCIL - Social role discovery from broadcast conversations
  • HRI - Human-robot interaction using audio and video inputs.
  • GALE - DARPA project, information distillation from multi-lingual textual and audio sources.
  • RT07s - NIST RT07s Meeting Recognition Evaluation, Spring 2007.
  • RT09s - NIST RT09s Meeting Recognition Evaluation, Spring 2009.
  • MeetingsEval06 - NIST ~RT06s Meeting Recognition Evaluation, Spring 2006.
  • MLP-based Features - The features formerly known as \"Novel Approaches\"
  • SpeakerRecognition - Speaker Recognition
  • SpeakerSegmentation - Research on diarization (speaker segmentation)
  • OverlappedSpeech - Research on overlapped speech handling for diarization
  • SmartWeb - German project for natural language interfaces to 2006 World Cup information
  • AMI/AMIDA - Augmented Multi-party Interaction (with Distant Access). April 2008 Report
  • Potential projects - Potential projects for visitors, new group members, and EE225D students
  • MeetingsEval05 - Meetings Recognition Evaluation 2005
  • MeetingsEval04 - Meetings Recognition Evaluation 2004
  • Metaphrase - Generating different ways of saying things using FrameNet
  • Poster2008 - The ICSI Speech Group Poster for 2008
  • Emotion - Some random thoughts on collecting a corpus of emotional speech from World of Warcraft (a massive multiplayer game).
  • Laughter Segmentation - A system which segments the audio into laughter and non-laughter regions.
  • Appscio - ICSI internal staging area for the Appscio Open ASR project.
  • Heart Challenge - Classifying Heart Sounds Challenge
  • RandomIdeas - Dumping ground for random ideas
  • PI Proposal Page - Page to keep track of current/pending/upcoming proposals
  • NSF Proposal Writing for Dummies - A quick and dirty tutorial to check your next idea against
  • Multimedia Privacy Brave New Idea - Brainstorming on a paper for the Brave New Ideas session of ACM Multimedia