AI Paper Atlas

Research map / Video / Action recognition

Task Graph Maximum Likelihood Estimation for Procedural Activity Understanding in Egocentric Videos

Luigi Seminara, Giovanni Maria Farinella, Antonino Furnari

TPAMI 2026 · Video › Action recognition · approach: Action anticipation and latent actions · 1 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
learn task graph by gradient-based maximum likelihood (DO) instead of MSG2 heuristicstask graph generation · CaptainCook4DF1: 73.3 -> 87.8Improved
DO instead of Count-Basedtask graph generation · EgoPERF1: 80.8 -> 91.0Improved
DO instead of Llama-3.1-405B-Instructtask graph generation · EgoProceLF1: 58.7 -> 72.3Improved
use feature-based TGT-text instead of direct adjacency optimization (DO)task graph generation · CaptainCook4DF1: 87.8 -> 80.8Hurt
use TGT-text instead of DO for downstream Ego-Exo4D optional keystep (instance-level)procedure understanding · Ego-Exo4D ValOptional keysteps: 74.52 -> 75.56Mixed
use DO instead of TGT-text for procedural mistakes (instance-level)procedure understanding · Ego-Exo4D ValProc. Mistakes: 78.83 -> 84.52Improved
learned graph (TGT-text, keystep prediction) instead of graph-based baseline at procedure levelprocedure understanding · Ego-Exo4D ValPrev. Keysteps: 64.56 -> 71.37Improved
use explicit DO task graph instead of implicit PREGO for mistake detection (GT actions)online mistake detection · Assembly101-Oaverage F1: 39.4 -> 75.9Improved
use DO instead of Count-Based on EPIC-Tent-O (GT actions)online mistake detection · EPIC-Tent-Oaverage F1: 57.7 -> 58.3Improved
use TGT-text instead of DO with more descriptive step namesonline mistake detection · EPIC-Tent-Oaverage F1: 58.3 -> 64.1Improved
replace ground-truth actions with predicted actions for DOonline mistake detection · Assembly101-Oaverage F1: 75.9 -> 53.5Hurt
DO mistake-class F1 with predicted actions vs MSG2online mistake detection · Assembly101-OMistake F1: 33.2 -> 28.1Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy