Hauptseite

LISTEN RWTH Workshop / Summer School 2018 in Bonn, Germany

The EU project LISTEN (http://www.listen-project.eu/) is concerned with various aspects of far-field large-vocabulary speech recognition. As part of this project, we will organize a three-day workshop in Bonn/Germany, 17-19 July 2018 (Tuesday-Thursday).

The workshop intends to bring together experts from academic and industrial research in this fast developing area of speech recognition. The workshop will include regular talks and keynote talks. In total, we expect an audience of about 50 researchers.

Venue and time

Venue: Ameron Hotel Koenigshof Bonn, Germany (directly located on the river Rhine; 20 km distance from airport Cologne/Bonn; 1 km from train station Bonn; 30 km from train station Cologne)

(Tentative) Program

   - Tuesday, July 17:    09:00-18:00 technical session,
   - Wednesday, July 18:  09:00-13:00 technical session,
                          14:00-18:00 social event,
   - Thursday, July 19:   09:00-17:00 technical session.

Keynotes

Confirmed keynote talks with (tentative) titles (45 minutes + questions/discussions):

Industry:

• Alex Acero, Apple, USA: Our work on HomePod.
• Jasha Droppo, Microsoft, USA: The Microsoft 2017 Conversational Speech Recognition System.
• Erik McDermott, Google, USA: Next generation acoustic models for automatic speech recognition.
• Björn Hoffmeister, Amazon, USA: tba.

Academia:

• Chin-Hui Lee, Georgia Tech, USA: Two-Stage Enhancement of Noisy and Reverberant Microphone Array Speech or Automatic Speech Recognition Systems Already Trained in Clean Conditions.
• Jon Barker, U Sheffield, UK: The CHiME-5 challenge for far-field conversational speech recognition in domestic environments.
• Satoshi Nakamura, NAIST, Japan: Towards listening while speaking: speech chain by deep learning.
• Reinhold Häb-Umbach, U Paderborn, Germany: Neural Network Supported Acoustic Beamforming and Source Separation for ASR.

Please scroll to the end to find the list of regular talks.

Travel

We encourage the workshop participants to arrive on Monday and to leave after the end of the last session.

By plane

Several airports can be used to travel to Bonn:

• Cologne/Bonn (Köln/Bonn, CGN): 23 km to the venue by car, 25 minutes with bus SB60 from CGN to Bonn main station (4 stops) then 1 km walking or taxi
• Dusseldorf (Düsseldorf, DUS): 83 km by car, about 1h50 with public transport
• Frankfurt (FRA): 154 km by car, about 1h15 - 2h00 with public transport

if you need assistance with your travel planning, our secretariat will be happy helping you (+49 241 80-21601, e-mail sek@i6.informatik.rwth-aachen.de).

Watch out: There are two airports which have names similar to some of the above but lie badly remote in the countryside, notably Frankfurt-Hahn (HHN) and Düsseldorf Weeze (NRN; also Airport Weeze, Verkehrsflughafen Niederrhein). If you end up there, we will have a good laugh at the workshop.

By train

The hotel is in walking distance to Bonn main station (Bonn Hbf). Bonn is typically reached via Cologne (Köln).

Booking a hotel room

(This part of the website is work in progress.) The venue is also the hotel where the (almost) all workshop participants will stay. (If you intend stay at an other place, pls. inform our secretariat.)

• Room rate: 129.00 EUR per night for a room "smart standard", which is either single room or double room with single use, incl. breakfast buffet.
  (For use with two people, a double room is 149.00 EUR per night incl. breakfast buffet.)

• Check-in: from 15:00 (3:00 PM). Important note: If you are likely to check in after 18:00 on arrival day, pls. inform the hotel about it.

• Check-out: until 12:00 (noon)

Organizing Committee

• Hermann Ney, Volker Steinbiss (RWTH)
• Athanasios Mouchtaris (FORTH)
• Volker Fischer (EML)
• Enrico Gianotti (Cedat85)

Regular talks from LISTEN consortium

Confirmed regular talks with (tentative) titles (15 minutes + 5 minutes questions/discussions) (not in temporal order and incomplete yet):

• Enrico Gianotti or other, Cedat, Italy: Applications of the technology.
• Volker Fischer / others, EML, Germany: The EML 2018 contribution to the LISTEN project #1.
• Volker Fischer / others, EML, Germany: The EML 2018 contribution to the LISTEN project #2 .
• Nikos Stefanakis , FORTH, Greece: (Microphone arrays).
• Tobias Menne, RWTH, Germany: tbd.
• NN, RWTH, Germany: tbd.

Regular talks from workshop participants

Confirmed regular talks with (tentative) titles (15 minutes + 5 minutes questions/discussions) (not in temporal order and incomplete yet):

• Alfons Juan and/or Albert Sanchis, UPV, Spain: Confidence measures.
• NN, UPV, Spain: tbd.
• Lauréline Pérotin, INRIA Nancy, France: Multichannel RNN-based separation of overlapping speech.
• Sunit Sivasankaran, INRIA Nancy, France: Reverberation-robust ASR, etc., see mail Emmanuel Vincent 29.03.2018.
• Petros Maragos or Makis Potamianos, U Athens, Greece: tbd.
• Honza Cernocky, U Brno, Czech Republic: Collection of re-transmitted data and impulse responses and remote ASR and speaker verification.
• Florian Metze, CMU, USA: Grounded Sequence to Sequence Transduction (multi-modal speech recognition).
• Sanjeev Khudanpur, JHU, USA: tbd.
• Marcel Katz or Patrick Naylor, Nuance, UK: tbd.
• Tim Fingscheidt, TU Braunschweig, Germany: Model fusion for phoneme recognition via turbo principle.
• Rainer Martin / others, U Bochum, Germany: Binaural source localization and separation.
• Rainer Martin / others, U Bochum, Germany: Aspects related to acoustic sensor networks.
• NN, KIT, Germany: tbd.
• Drude, U Paderborn, Germany: tbd.

(internal-only)