Welcome to the Data-Driven Decision Making Lab (D3M)
Mission:
We research fundamental techniques and methodology from the fields of (Generative) AI, Machine Learning, Recommendation, and Information Retrieval to transform rich sources of data into optimal and actionable decisions.
Research Areas:
Machine Learning and Large-scale Data Analysis, Artificial Intelligence (AI), Information Retrieval, Social Media, Recommender Systems, Sequential Decision Optimization, Operations Research, Smart Cities Applications.
Lab News:
- Jan 15, 2024: D3M presented a AAAI-24 Lab and Tutorial on the Python PyRDDLGym rewrite of the RDDL MDP modeling software stack that includes the state-of-the-art JaxPlan solver for planning via autodifferentiation.
- Aug 3, 2023: Check out our pure LLM-based conversational recommendation system LLM-ConvRec including a Colab demo for restaurant recommendation that provides a fluent and review-informed interactive experience.
- Jan 9, 2023: Check out an outstanding summary overview of our IPM 2023 article on Unintended Bias in Language Model-driven Conversational Recommendation written by D3M MASc graduate Tianshu (Tina) Shen.
- June 17, 2020: Congratulations to D3M MASc student Zheda (Marco) Mai who won 1st place in the CVPR 2020 CLVISION Challenge with his entry Batch-level Experience Replay with Review.
- Feb 28, 2020: Professor Sanner has received a Google Faculty Research Award in Machine Learning and Data Mining to pursue research on Deep Conversational Recommender Systems.
- May 25, 2018: Zhijiang (Tony) Ye’s paper with Buser Say entitled Symbolic Bucket Elimination for Piecewise Continuous Constrained Optimization has received the Student Paper Award at CPAIOR-18.
- June 28, 2017: Professors Scott Sanner and William O’Brien (Carleton) and D3M student Brent Huchuk have been awarded an NSERC CRD to pursue research on machine learning for residential HVAC with Ecobee, Inc.
- Dec 29, 2016: Iain Guilliard has won the 2016 Kikuchi-Karlaftis Best Paper Award of the Transport Research Board for his paper A Non-homogeneous Time Mixed Integer LP Formulation for Traffic Signal Control (video).