NeurIPS 2020 : Competitions

Machine learning competitions have grown in popularity and impact over the last decade, emerging as an effective means to advance the state of the art by posing well-structured, relevant, and challenging problems to the community at large. Motivated by a reward or merely the satisfaction of seeing their machine learning algorithm reach the top of a leaderboard, practitioners innovate, improve, and tune their approach before evaluating on a held-out dataset or environment. The competition track of NeurIPS has matured in 2020, its fourth year, with a considerable increase in both the number of challenges and the diversity of domains and topics. A total of 16 competitions are featured this year as part of the track, with 8 competitions associated to each of the two days. The list of competitions that are part of the program are available here: https://neurips.cc/Conferences/2020/CompetitionTrack

Competition Track Friday

Competition Track Saturday

Competition Track Friday

Hugo Jair Escalante, Katja Hofmann

2020-12-11T08:00:00-08:00 - 2020-12-11T17:45:00-08:00

Join Zoom

Toggle Abstract

Abstract: First session for the competition program at NeurIPS2020.

Machine learning competitions have grown in popularity and impact over the last decade, emerging as an effective means to advance the state of the art by posing well-structured, relevant, and challenging problems to the community at large. Motivated by a reward or merely the satisfaction of seeing their machine learning algorithm reach the top of a leaderboard, practitioners innovate, improve, and tune their approach before evaluating on a held-out dataset or environment. The competition track of NeurIPS has matured in 2020, its fourth year, with a considerable increase in both the number of challenges and the diversity of domains and topics. A total of 16 competitions are featured this year as part of the track, with 8 competitions associated to each of the two days. The list of competitions that ar part of the program are available here:

https://neurips.cc/Conferences/2020/CompetitionTrack

Live Video

Chat is not available.

Schedule

2020-12-11T08:00:00-08:00 - 2020-12-11T08:15:00-08:00

Opening - Competition Track Session

Katja Hofmann, Hugo Jair Escalante

Competition Track Friday

Hugo Jair Escalante, Katja Hofmann

Live Video

Chat is not available.

Schedule

Opening - Competition Track Session

3D+texture garment reconstruction challenge design (data, metrics, tracks, etc)

3D+texture garment reconstruction challenge results

Opening the L2RPN challenge @ NeurIPS2020

Winning the L2RPN challenge

A L2RPN Winning approach

The Best L2RPN wining approach

L2RPN Post Challenge open questions

Closing and ceremony award

Introducing the Hide-and-Seek privacy challenge

The importance of synthetic data

Synthetic data in the healthcare setting

What we learned from the Hide-and-Seek privacy challenge

Closing remarks

Background on black box optimization (BBO)

BBO challenge platform with Valohai

Spotlight for 1st place (BBO challenge)

Spotlight for 2nd place (BBO challenge)

Spotlight for 3rd place (BBO challenge)

Opening the SpaceNet 7 Challenge @ NeurIPS2020

Introduction to SpaceNet

The SpaceNet 7 Dataset

The SpaceNet 7 Metric

The Winners of SpaceNet 7

SpaceNet 7 Closing and Future Plans

introduction to the 2020 NeurIPS education challenge

Competition overview: motivation, impact, dataset, tasks

Competition results and insights

Beyond the competition: what's next?

Q&A and discussion

Traffic Map Movies - An Introduction to the Traffic4cast Challenge

The Traffic4cast Competition Design and Data

The Best Traffic4Cast Submissions

1st prize: Utilizing UNet for the Future Traffic Map Prediction - Traffic4cast highlight talk

2nd prize: TLab: Traffic Map Movie Forecasting Based on HR-NET - Traffic4cast highlight talk

3rd prize: Towards Good Practices of U-Net for Traffic Forecasting - Traffic4cast highlight talk

Graph Ensemble Net and the Importance of Feature & Loss Function Design for Traffic Prediction - Traffic4cast highlight talk

Uncertainty Intervals for Graph-based Spatio-Temporal Traffic Prediction - Traffic4cast highlight talk

Traffic4cast Award Ceremony, Outlook, and Follow Up Challenges

The Hateful Memes Challenge: Competition Overview

The Hateful Memes Challenge: Live award ceremony and winner presentations

Competition Track Saturday

Hugo Jair Escalante, Katja Hofmann

Live Video

Chat is not available.

Schedule

Introducing EfficientQA: Open domain question answering with memory constraints as a testbed for language understanding and knowledge representations

Track winner presentations:

Showdown against trivia experts

Predicting Generalization in Deep Learning (PGDL): Opening remark

Keynote speech: Sanjeev Arora (PGDL)

Introduction to winning team (PGDL)

Winning team presentation: On Representations and Generalization (PGDL)

Introduction to Runner up 1 (PGDL)

Runner up presentation: Robustness to Augmentations as a Generalization Metric (PGDL)

Introduction to Runner up 2 (PGDL)

Runner up presentation: Ranking generalization via smoothness of latent graphs (PGDL)

Closing remark (PGDL)

Openning and design of the INTERPRET challenge @ NeurIPS2020

Winner talks of INTERPRET challenge

Analysis, research opportunities and closing of INTERPRET challenge

NLC2CMD Competition Organizers: Introduction, Problem Description, CLAI

NLC2CMD Competition Keynote: Tellina

NLC2CMD Competition Organizers: Metrics, Data, Tracks

NLC2CMD Competition Participant Team: AINixClaiSimple

NLC2CMD Competition Participant Team: coinse-team

NLC2CMD Competition Participant Team: AICore

NLC2CMD Competition Participant Team: magnum

NLC2CMD Competition Participant Team: Hubris

NLC2CMD Competition Participant Team: jb

NLC2CMD Competition Organizers: Results (Live)

Introduction to AIDO

Short Scientific Talk (AIDO)

Advanced Perception League

Intro to Urban League (includes highlights from semifinals)