Papers › AssembleNet++: Assembling Modality Representations via Attention Connections

AssembleNet++: Assembling Modality Representations via Attention Connections

18 Aug 2020arXiv:2008.08072archive 2025-07-28

Michael S. Ryoo, AJ Piergiovanni, Juhana Kangaspunta, Anelia Angelova

We create a family of powerful video models which are able to: (i) learn interactions between semantic object information and raw appearance and motion features, and (ii) deploy attention in order to better learn the importance of features at each convolutional block of the network. A new network component named peer-attention is introduced, which dynamically learns the attention weights using another block or input modality. Even without pre-training, our models outperform the previous work on standard public activity recognition datasets with continuous videos, establishing new state-of-the-art. We also confirm that our findings of having neural connections from the object modality and the use of peer-attention is generally applicable for different existing architectures, improving their performances. We name our model explicitly as AssembleNet++. The code will be available at: https://sites.google.com/corp/view/assemblenet/

PaperPDFCode

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Action ClassificationActivity Recognition

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Action Classification Charades AssembleNet++ 50 MAP 59.8 #5 of 49 Archive leaderboard report
Action Classification Charades AssembleNet++ 50 without object MAP 54.98 #9 of 49 Archive leaderboard report
Action Classification Toyota Smarthome dataset AssembleNet++ CS 63.6 #5 of 13 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

Peer-attention

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections