Adding a backward prediction task to LLM training improves long-term action anticipation on Ego4D, lowering edit distance for predicted action sequences.
Multiscale vision transformers
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
Bidirectional Action Sequence Learning for Long-term Action Anticipation with Large Language Models
Adding a backward prediction task to LLM training improves long-term action anticipation on Ego4D, lowering edit distance for predicted action sequences.