MLab

ACMLab is Stanford's premier machine learning club. Its goal is to teach anyone with basic CS experience machine learning. After an intensive ramp-up workshop in the fall, members work on publishing papers at top ML conferences and workshops. We have published 6 workshop papers so far at top conferences and workshops such as ACL and ICLR. Alumni have gone on to Google AI, Stanford ML Group, Stanford NLP Group, and VMWare.

Board

Ethan Chi

2022

Co-Director

Jillian Tang

2022

Co-Director

Gordon Chi

2023

Langston Nashold

2023

Yasmine Mitchell

2023

William Ellsworth

2023

Recent Projects

SemEVAL

We submitted to the Workshop on Semantic Evaluation’s Task 1 (lexical complexity modelling) and Task 8 (automatically extracting measurements from scientific text). Our teams performed competitively on both tasks, including second place in one of the Task 8 subcategories. Our task description papers will appear at SemEval at ACL 2021.

Google BIG-Bench

Members proposed 4 tasks to be used in Google's BIG-Bench challenge. The purpose of this challenge was to create a collaborative benchmark for enourmous language models like GPT-3. MLab submitted tasks about temporal sequences, logic puzzles, sarcasm, and IPA translation.

VQA

We are currently preparing a submission on the ChartQA workshop at CVPR 2021, aiming to automatically parse structured information from diverse chart-based visual representations.

Recent Publications

An IPA Translation Task for Probing Large Language Models.

O Zhang, EA Chi

The Workshop on Enormous Language Models @ ICLR 2021.


SNARKS: Single-Edit Non-Contextual Adversarial Knowledge-agnostic Sarcasm.

R Chi, I Ng, EA Chi, T Chang

The Workshop on Enormous Language Models @ ICLR 2021.


A Logic Puzzles Task for Probing Large Language Models.

R Garg, W Zhang, R Sikand, J Kim, EA Chi, J Tang

The Workshop on Enormous Language Models @ ICLR 2021.


A Temporal Sequences Task for Probing Large Language Models.

H Kim, J Zheng, EA Chi†

The Workshop on Enormous Language Models @ ICLR 2021.


Stanford MLab at SemEval Shared Task 8: 48 Hours is All You Need.

P Liu, NS Iyer, E Rozi, EA Chi†

SemEval 2021 @ ACL-IJCNLP 2021.


Stanford MLab at SemEval Shared Task 1: Tree-Based Methods for Lexical Complexity Prediction.

E Rozi*, NS Iyer*, E Choe, G Chi, KJ Lee, K Liu, P Liu, Z Lack, J Tang†, EA Chi†

SemEval 2021 @ ACL-IJCNLP 2021.