The 2013 IEEE Workshop on Automatic Speech Recognition and Understanding was held in Olomouc, Czech Republic from December 8th to 12th. This was probably the best conference I have attended in my brief academic career. I was very impressed with the organization and the quality of research and speakers.
The plan for the conference was different than any I have seen before. All of the approximately 85 accepted papers were presented as posters. Since the number of posters was relatively small, it was possible to hang all posters in a single room for the duration of the workshop. There were no poster sessions. Instead, authors were requested to be around their posters at specific times throughout the week. People were encouraged to also set up meetings at other times to further discuss their research. I was skeptical of this approach prior to attending, but it turned out great. By the end of the week I had plenty of time to see any poster I was interested in, and talk to the authors.
In the time between poster sessions, we were treated to invited talks. No talk overlapped in time, so it was possible to see every talk in its entirety. Most of the talks were about broader ideas, so it was possible to walk away with a lot of new information. I think these types of talks can be more useful than highly technical talks about a very narrow research topic.
I have already written a series of posts about the workshop, but I wanted to link them together again in one place.
Daily recap of invited talks:
Day 1: Neural Networks
Day 2: Limited Resource
Day 3: Applications
Day 4: What is Wrong with ASR?
Overview of a few interesting papers, and shameless promotion of my own work.
Between Zero and One 