We are live streaming the event! Notice that Day 2 and Day 1 use different links.
Toyota Technological Institute at Chicago
September 4- September 5
In the last few years, foundation models for speech and audio have advanced rapidly. Today’s pre-trained speech and audio encoders allow for quick adaptation to a variety of tasks, such as speech recognition, speech synthesis, speaker identification, spoken language understanding, audio event detection, and music understanding and generation. Recent work has also been developing general-purpose language models for speech and audio, which promise to perform arbitrary tasks in a single model, similarly to LLMs for text. This workshop brings together researchers and practitioners at the forefront of speech and audio foundation model development to discuss foundational principles, emerging trends, and current challenges. Topics include model architectures and pre-training strategies, multilingual and multimodal models, and evaluation benchmarks. The workshop will include speakers and participants across academia and industry to foster an exchange of ideas and help chart a path for the emerging crop of robust, general-purpose speech and audio foundation models.
Venue: TTIC, Chicago, USA (September 4 - September 5)
Registration is free here
We invite abstracts describing published work and work in progress for poster presentation at the workshop. All topics involving speech and audio foundation models, including multimodal models that involve text and/or other modalities, are considered relevant.
Abstracts are non-archival and no proceedings will be published. Submissions should include titles and abstracts, but not author names. Abstracts may be submitted as either text, normally less than 500 words, or along with a PDF of formatted text, tables, and figures. There is no specific format or template for PDF submissions. When submitting the work, you can opt to also present the work with a short highlight presentation (5-10 mins). We will select the presentations based on schedule availabilities.
Please make your submission before August 5, 11:59 pm AOE.
Hung-yi Lee
National Taiwan University (NTU)
Noah Smith
University of Washington & AI2
Yossi Adi
Hebrew University FAIR, Meta
David Harwath
UT Austin
Wei-Ning Hsu
Meta Superintelligence Lab
Tatiana Likhomanenko
Apple
Mingqiu Wang
Google Deepmind
Neil Zeghidour
Kyutai
9:15 - 9:50 Registration & Breakfast
9:50 - 10:00 Opening Remarks
10:00 - 11:00 Keynote Talk: Hung-yi Lee
11:00 - 11:10 Break
11:10 - 11:50 Invited Talk: Neil Zeghidour
11:50 - 12:30 Invited Talk: Mingqiu Wang
12:30 - 13:05 Lightning Talks Session 1
13:05 - 14:40 Lunch & Poster Session 1
14:40 - 15:20 Invited Talk: Yossi Adi
15:20 - 16:00 Invited Talk: Wei-Ning Hsu
16:00 - 16:10 Break
16:10 - 17:30 Nuts & Bolts Session 1: Democratization of Speech Foundation Models
Led by: Shinji Watanabe
9:00 - 9:30 Breakfast
9:30 - 10:30 Keynote Talk: Noah Smith
10:30 - 10:40 Break
10:40 - 11:20 Invited Talk: David Harwath
11:20 - 12:00 Invited Talk: Tatiana Likhomanenko
12:00 - 12:40 Lightning Talks Session 2
12:40 - 14:15 Lunch & Poster Session 2
14:15 – 15:00 Nuts & Bolts Session 2: Evaluation
Led by: Hung-yi Lee
15:00 - 15:10 Break, panel setup
15:10 - 16:10 Nuts & Bolts Session 3: The Industry Experience
Panel: industry speakers
16:10 - Close
For more detailed schedule and talk information, check Schedule & Talks.
There are several hotels nearby the University of Chicago campus in Hyde Park, where TTIC is located. These include:
The Study (a five minute walk from TTIC)
There are also several reasonable options in downtown Chicago, such as Hotel Felix. To search for more options, you can check nearby hotels.
Parking:
Free parking is available in the commuter parking lot at 60th St. and Stony Island Ave. and free street parking on many streets near TTIC (just beware of "permit parking" and "street cleaning" signs!). Parking can be found on 61st Street (between Woodlawn Ave and Blackstone Ave), on Dorchester Street (between 60th and 61st Streets).
Registration is free. Please sign up before August 13 below if you are interested to attend the workshop. To present your work as a poster or talk, please also complete the poster submission below. We are awarding a limited number of small grants to attendees including students/junior researchers/presenters who are traveling from outside the Chicagoland area. Please fill out this form to indicate your need.