Methods › General › Attention Modules › Feedback Memory
Feedback Memory
Introduced by Angela Fan et al. in Addressing Some Limitations of Transformers with Feedback Memory
archive 2025-07-28 Description, source and code snippet are the archive's method entry.
Feedback Memory is a type of attention module used in the Feedback Transformer architecture. It allows a transformer to to use the most abstract representations from the past directly as inputs for the current timestep. This means that the model does not form its representation in parallel, but sequentially token by token. More precisely, we replace the context inputs to attention modules with memory vectors that are computed over the past, i.e.:
zˡₜ = Attn(xˡₜ, [m_(t-τ), …, mₜ₋₁])
where a memory vector mₜ is computed by summing the representations of each layer at the t-th time step:
mₜ = ∑ᴸₗ₌₀Softmax(wˡ)xₜˡ
where wˡ are learnable scalar parameters. Here l = 0 corresponds to token embeddings. The weighting of different layers by a softmax output gives the model more flexibility as it can average them or select one of them. This modification of the self-attention input adapts the computation of the Transformer from parallel to sequential, summarized in the Figure. Indeed, it gives the ability to formulate the representation xˡₜ₊₁ based on past representations from any layer l′, while in a standard Transformer this is only true for l > l′. This change can be viewed as exposing all previous computations to all future computations, providing better representations of the input. Such capacity would allow much shallower models to capture the same level of abstraction as a deeper architecture.
In Syntology name not resolved (unknown_name). The lookup matched no method, paper or venue by this name; that asserts nothing about whether code exists.
Near matches the lookup lists, each a different name: Distilling Feedback into Memory-as-a-Tool; Resource-Efficient Iterative LLM-Based NAS with Feedback Memory; Memory-R2; TS-Memory; Memory-QA; Memory-V2V; OCR-Memory; Memory Lobe.
Syntology's have check for this name (free, it answers coverage only), read 2026-10-05; ask it again live. A level is the served verification level of that one answer; nothing here is summed across methods.
Papers archive 2025-07-28
archive papers tagged: 4 · with a code link: 3 · where Syntology ran a sample: 1 (1 with a run with no instrument failure, 0 where every run was a failure of Syntology's instrument) Syntology
Show: all tagged papersonly where code ran (1 of 4 tagged: 1 with a run with no instrument failure, 0 where every run was a failure of Syntology's instrument)
Page 1 of 1: papers 1 to 4 of 4, newest first by the archive's date (ties by slug), in archive order.
Repository counts are the archive's code-links table. A Syntology line states what Syntology ran from that paper's harvested code, as “N ran (of which C constructed an object rather than computing a result; K with no instrument failure: H honoured, V violated, P with no contract checked; I where Syntology's instrument failed) · U unverified”; the instrument figure counts failures of Syntology's instrument, not of the code. It is per sample and not a correctness claim. When the archive marks a repository official for the paper, the line starts with that repository's state (the archive's flag, not a verdict on who wrote the code; “community repositories only” when every sample that ran came from a community repository, “official: no sample here; runs from other or unrecorded repositories” when some came from a repository the paper names or has in its text, or from none recorded); hover it for the repositories the samples that ran came from.
-
SPIRe: Boosting LLM Inference Throughput with Speculative Decoding 8 Apr 2025 · 0 repositories · arXiv:2504.06419
-
Iterative Feedback Network for Unsupervised Point Cloud Registration 9 Jan 2024 · 1 repository · arXiv:2401.04357
-
Do You Know My Emotion? Emotion-Aware Strategy Recognition towards a Persuasive Dialogue System 24 Jun 2022 · 1 repository · arXiv:2206.12101
-
Addressing Some Limitations of Transformers with Feedback Memory 21 Feb 2020 · 4 repositories · arXiv:2002.09402Syntology official: no sample here; runs from other or unrecorded repositories · 3 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 0 honoured, 2 violated, 0 with no contract checked; 1 where Syntology's instrument failed) · 0 unverified (of 3 harvested samples) · 2 pointer-only (licence)
Tasks archive 2025-07-28
6 tasks the archive attaches to papers tagged with this method, by distinct papers. A task without a page in the catalog is plain text.
| Task | Papers |
|---|---|
| Language Modeling | 1 |
| Language Modelling | 1 |
| Machine Translation | 1 |
| Point Cloud Registration | 1 |
| Reinforcement Learning | 1 |
| Translation | 1 |
Usage over time archive 2025-07-28
Components: the archive holds no method-to-method composition, so PwC's Components table cannot be rebuilt; the Papers list carries no Results column for the same reason (the archive does not join its leaderboard rows to method tags).
Categories archive 2025-07-28
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