Project led by Vincent Cartillier (PhD student at Machine Learning at Georgia Tech) together with Zhile Ren (GT --> Apple), Neha Jain (GT --> FB), Stefan Lee Stefan Lee (Oregon State University), and Irfan Essa Irfan Essa (Machine Learning at Georgia Tech Georgia Tech School of Interactive Computing Georgia Tech Computing).

Advancing first-person perception with 2022 Ego4D challenge ai.facebook.com/blog/advancing…

Join the #Ego4D challenge, exploring the largest ever dataset of first-person video and five new research benchmarks: episodic memory, hands+objects, social, AV, forecasting. First round of the competition ends June 1 with results shared at #CVPR. ego4d-data.org/docs/challenge/

Are you planning to submit to any of the Ego4D Challenges this #CVPR2026 and be the first to win the #Ego4D open challenges? Deadline 1st of June Let us know which challenges you are considering/preparing for? ego4d-data.org/#challenges

A first in-person conference presentation at #CVPR2022 in New Orleans this week!🎉 We introduce a new task Episodic Memory Question Answering – accepted as an Oral presentation. Full details (and paper) here : samyak-268.github.io/emqa/

Thank you to everybody who showed up at the Oral talk / poster session! #CVPR2022

SLAIM: Robust Dense Neural SLAM for Online Tracking and Mapping vince_cartillier, Grant Schindler, Irfan Essa tl;dr: KL regularizer over ray termination distribution; image Gaussian pyramid filter (similar to hierarchical Lucas-Kanade optical flow)->NeRF arxiv.org/pdf/2404.11419…

Just received an invitation to review a paper for the PLOS ONE journal. I've only reviewed for CV/ML conferences. The paper is about human perception and cognition. Have you ever reviewed in other venues than traditional CS? Trying to see if I qualify for it/should do it!?

[CODE DROP] Ever wonder how to do robust NeRF-SLAM? We present SLAIM: a coarse-to-fine nerf-slam implementation in c++ and cuda kernels on top of Instant-NGP. code+project: vincentcartillier.github.io/slaim.html Come see us at CVPR - NRI workshop! Irfan Essa

World Models have gained significant momentum in the research community over the past few years. However, we still lack systematic approaches for evaluating them properly for downstream applications and making informed design decisions. We're organizing the WorldModelBench

Amazon is back with Nova 2.0, a substantial upgrade over prior Amazon Nova models and demonstrating particular strength in agentic capabilities Amazon has released Nova 2.0 Pro (Preview), its new flagship model; Nova 2.0 Lite, focused on speed and lower cost; and Nova 2.0 Omni,

Nova 2.0 Pro Preview demonstrates particular strength in agentic capabilities, scoring 93% on τ²-Bench Telecom and 80% on IFBench on medium and high reasoning budgets respectively (complete benchmarks for high reasoning coming soon). This places it amongst the leading models in