facebookresearch / facebookresearch/sparse-delta-memory
Doubt about the equation 4 of the paper SDM
Open
- Dominant language
- Python
- Stars
- 39
- Forks
- 3
- PR merge metrics
- No merged PRs in 30d
Description
Hello,
I was reading the paper and I got a bit confused by eq. 4. Shouldn't the term subtracted from $v$ be the weighted sum of $\bar{M}_t$ over all selected write indices, rather than just the single $\bar{M}_t$ term?
Thank you a lot for your time !
Contributor guide
Research direction
The issue names no files, tests, or entry points. Start by checking equation 4 in the Sparse Delta Memory paper against the reference implementation, then document whether the equation or implementation is correct and explain the role of the selected write indices.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- documentation
- Issue type
- Documentation
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Active
- Clarity
- Needs clarification
- Newbie friendliness
- 30/100