AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

Token Turing Machines

Michael S. Ryoo, Keerthana Gopalakrishnan, Kumara Kahatapitiya, Ted Xiao, Kanishka Rao, Austin Stone, Yao Lu, Julian Ibarz, Anurag Arnab

CVPR 2023 · Architectures and efficiency › Attention and transformers · approach: Efficient language transformers · 13 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
Add TTM to ViViT-BViViT-B · Charades v1 localize validationmAP: 23.18 -> 26.34Improved
Add TTM to ViViT-LViViT-L · Charades v1 localize validationmAP: 26.01 -> 28.79Improved
Use TTM-Transformer with 3136 input tokens rather than a temporal Transformer with 3360 tokensViViT-B · Charades v1 localize validationmAP: 25.53 -> 26.34Improved
Add TTM per videoViViT-B · AVA v2.2mAP: 25.2 -> 27.9Improved
Add TTM per boxViViT-B · AVA v2.2mAP: 25.2 -> 31.5Improved
Use TTM memory for policy historyBC-Zero EfficientNet · Real-robot evaluationtask success rate: 80.08 -> 89.26Improved
Replace Transformer processing unit with MLP (ablation)TTM · Charades v1 localize validationmAP: 26.34 -> 23.34Hurt
Replace Transformer processing unit with MLPMixer (ablation)TTM · Charades v1 localize validationmAP: 26.34 -> 26.14No clear effect
Replace MLP summarization with latent-query summarization (ablation)TTM · Charades v1 localize validationmAP: 26.34 -> 26.75Improved
Replace learned token summarization with pooling (ablation)TTM · Charades v1 localize validationmAP: 26.34 -> 25.75Hurt
Replace bounded memory update with concatenation of observed input tokens (ablation)TTM · Charades v1 localize validationmAP: 26.34 -> 20.97Hurt
Replace token summarization write with erase-and-add memory writes (ablation)TTM · Charades v1 localize validationmAP: 26.34 -> 25.86Hurt

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