Affinity Workshop
New In Machine Learning (NewInML)
Alice Lacan · Mélina Verger
Room 301 - 303
Following the previous NewInML workshop editions, our goal is to welcome newcomers in the community and provide them with some guidance to contribute to Machine Learning research fully and effectively.
We intend to host presentations appealing to ICML audiences eager to learn how to conduct their research from experienced researchers. The following topics will be addressed:
- Communicating your results
- Collaborations with ML researchers
- Coding best practices
Attendees who submit an extended abstract will get peer-reviewed and best selected abstracts will be presented by their authors. This year, we are also hosting a presentation on academic writing support tailored on the submissions. It is your opportunity to refine your academic writing skills and boost your chances to see your paper published at the next ICML conference.
Schedule
|
Mon 6:00 a.m. - 6:05 a.m.
|
Opening address
(
Introduction
)
SlidesLive Video |
Alice Lacan · Mélina Verger 🔗 |
|
Mon 6:05 a.m. - 6:45 a.m.
|
Invited Talk #1 - How to communicate your results
(
Invited Talk
)
link
SlidesLive Video |
Andrew Fitzgibbon 🔗 |
|
Mon 6:45 a.m. - 6:55 a.m.
|
Invited Talk #1 - Q&A
(
Discussion
)
|
🔗 |
|
Mon 6:55 a.m. - 7:35 a.m.
|
Invited Talk #2 - Collaborations with ML researchers
(
Invited Talk
)
link
SlidesLive Video |
Rosanne Liu 🔗 |
|
Mon 7:35 a.m. - 7:45 a.m.
|
Invited Talk #2 - Q&A
(
Discussion
)
|
🔗 |
|
Mon 7:45 a.m. - 8:15 a.m.
|
Coffee Break
(
Break
)
|
🔗 |
|
Mon 8:15 a.m. - 8:55 a.m.
|
Invited Talk #3 - Coding best practices
(
Invited Talk
)
link
SlidesLive Video |
Jill-Jênn Vie 🔗 |
|
Mon 8:55 a.m. - 9:05 a.m.
|
Invited Talk #3 - Q&A
(
Discussion
)
|
🔗 |
|
Mon 9:05 a.m. - 9:10 a.m.
|
Selected Paper #1 - FedControl: When Control Theory Meets Federated Learning
(
Talk
)
SlidesLive Video |
Alexandre Duplessis 🔗 |
|
Mon 9:10 a.m. - 9:15 a.m.
|
Selected Paper #1 - Q&A
(
Discussion
)
|
Alexandre Duplessis 🔗 |
|
Mon 9:15 a.m. - 9:20 a.m.
|
Selected Paper #2 - Test-Time Adaptation with Principal Component Analysis
(
Talk
)
SlidesLive Video |
Thomas Cordier 🔗 |
|
Mon 9:20 a.m. - 9:25 a.m.
|
Selected Paper #2 - Q&A
(
Discussion
)
|
Thomas Cordier 🔗 |
|
Mon 9:25 a.m. - 9:30 a.m.
|
Selected Paper #3 - Light Weight Character and Shape Recognition for Autonomous Drones
(
Talk
)
SlidesLive Video |
Neetigya Poddar 🔗 |
|
Mon 9:30 a.m. - 9:35 a.m.
|
Selected Paper #3 - Q&A
(
Discussion
)
|
🔗 |
|
Mon 9:35 a.m. - 9:40 a.m.
|
Selected Paper #4 - A Compact Transformer-based Classifier with Selected Hybrid Features from Different Patch Sequences for Image Classification
(
Talk
)
SlidesLive Video |
Luna Zhang 🔗 |
|
Mon 9:40 a.m. - 9:45 a.m.
|
Selected Paper #4 - Q&A
(
Discussion
)
|
Luna Zhang 🔗 |
|
Mon 9:45 a.m. - 9:50 a.m.
|
Selected Paper #5 - An Efficient Modern Baseline for FloodNet VQA
(
Talk
)
SlidesLive Video |
Aditya Kane 🔗 |
|
Mon 9:50 a.m. - 9:55 a.m.
|
Selected Paper #5 - Q&A
(
Discussion
)
|
Aditya Kane 🔗 |
|
Mon 9:55 a.m. - 10:00 a.m.
|
Selected Paper #6 - Generating Synthetic Population
(
Talk
)
SlidesLive Video |
Bhavesh Neekhra 🔗 |
|
Mon 10:00 a.m. - 10:05 a.m.
|
Selected Paper #6 - Q&A
(
Discussion
)
|
Bhavesh Neekhra 🔗 |
|
Mon 10:15 a.m. - 12:00 p.m.
|
Lunch Break
(
Break
)
|
🔗 |
|
Mon 12:00 p.m. - 1:20 p.m.
|
Writing Effectively in ML
(
Discussion
)
SlidesLive Video |
Réka Mihálka 🔗 |
|
Mon 1:20 p.m. - 1:25 p.m.
|
Data Challenge Presentation: Cross-Domain MetaDL
(
Sponsor Data Challenge presentation
)
link
SlidesLive Video The new Cross-Domain MetaDL challenge is part of the ChaLearn meta-learning series. It has a special league for New in ML participants (with prizes and certificates) and a detailed tutorial with no prerequisites (i.e., no previous meta-learning knowledge required). The competition is part of the NeurIPS'22 program, and the winners will be invited to co-author the analysis paper with the organizers to appear in PMLR. The focus is on "cross-domain" meta-learning, aiming at leveraging experience from previous tasks to solve new tasks efficiently. While our previous challenge addressed within-domain few-shot learning for N-way k-shot tasks (i.e., N class classification problems with k training examples), this challenge proposes "any-way" and "any-shot" tasks drawn from various domains (healthcare, ecology, biology, manufacturing, and others), chosen for their humanitarian and societal impact. Code submissions will be blind-tested on CodaLab, and the winners' code will be open-sourced. |
🔗 |
|
Mon 1:25 p.m. - 1:30 p.m.
|
Closing remarks
(
Closing remarks
)
|
Alice Lacan · Mélina Verger 🔗 |