Methods › General › Momentum Rules › Demon

Demon

16 papers tagged archive 2025-07-28

Introduced by John Chen et al. in Demon: Improved Neural Network Training with Momentum Decay

archive 2025-07-28 Description, source and code snippet are the archive's method entry.

Decaying Momentum, or Demon, is a stochastic optimizer motivated by decaying the total contribution of a gradient to all future updates. By decaying the momentum parameter, the total contribution of a gradient to all future updates is decayed. A particular gradient term gₜ contributes a total of η∑ᵢβⁱ of its "energy" to all future gradient updates, and this results in the geometric sum, ∑^∞ᵢ₌₁βⁱ = β∑^∞ᵢ₌₀βⁱ = β/(1-β). Decaying this sum results in the Demon algorithm. Letting βᵢₙᵢₜ be the initial β; then at the current step t with total T steps, the decay routine is given by solving the below for βₜ:

βₜ/(1-βₜ) = (1-t/T)βᵢₙᵢₜ/(1-βᵢₙᵢₜ)

Where (1-t/T) refers to the proportion of iterations remaining. Note that Demon typically requires no hyperparameter tuning as it is usually decayed to 0 or a small negative value at time T. Improved performance is observed by delaying the decaying. Demon can be applied to any gradient descent algorithm with a momentum parameter.

PaperSourceSee Code · JRC1995/DemonRangerOptimizer

In Syntology about another paper, arXiv:2308.04152: harvested code stored, below machine verification (unverified_code; harvested), served level V1; retrieve with get_harvested_code_for_paper(arxiv_id='2308.04152'). Repositories: 2 listed of 2 in the graph, strongest link: official. The lookup matched a paper by its title, arXiv:2308.04152, not a method entry. About another paper: this answer is about arXiv:2308.04152, while the archive credits arXiv:1910.04952.

From the lookup's note on this name: The same name also belongs to the paper “DeMoN: Depth and Motion Network for Learning Monocular Stereo” (arXiv:1612.02401, 751 citations), which the lookup finds cited far more than the paper this answer is about (98 citations), so this answer may be about a different sense of the name. For arXiv:1612.02401: its page on this site · ask the lookup about it.

Origin paper of this answer, arXiv:2308.04152 (its page on this site · its page on Syntology): Syntology's run record for its harvested code, 10 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 0 honoured, 1 violated, 1 with no contract checked; 8 where Syntology's instrument failed) · 2 unverified (of 12 harvested samples) · 1 pointer-only (licence). The instrument figure counts failures of Syntology's instrument, not of the code. Ran means executed on a synthesized input, not that the code is correct.

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.

What each level means, quoted from Syntology's server card: V0 “imports”; V1 “executes against a synthesized fixture and returns a value of the recorded type”; V2 “two implementations GENERATED from the paper's signature and spec agree numerically (1e-6) on fixture inputs” “neither of the two ever read the authors' own code, so V2 measures how unambiguously the paper specifies the method, not fidelity to the authors' implementation (measured 2026-09-11, commit 75eaa77: cross_checked_by is a second LLM on 2,628 of 2,831 served samples)”; V3 “paper-derived property tests pass on both of those same two generations.” The card adds: “No rung on this ladder compares the served code against the authors' implementation, so treat every level as evidence about runnability and internal agreement, not about matching the paper's results.” A level quoted in two parts is one definition the card writes as two clauses joined by a double hyphen, each clause quoted on its own (how these levels are quoted).

Papers archive 2025-07-28

archive papers tagged: 16 · with a code link: 5 · where Syntology ran a sample: 2 (0 with a run with no instrument failure, 2 where every run was a failure of Syntology's instrument) Syntology

Show: all tagged papersonly where code ran (2 of 16 tagged: 0 with a run with no instrument failure, 2 where every run was a failure of Syntology's instrument)

Page 1 of 1: papers 1 to 16 of 16, 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.

Tasks archive 2025-07-28

18 tasks the archive attaches to papers tagged with this method, by distinct papers. A task without a page in the catalog is plain text.

TaskPapers
Graph Neural Network2
Pose Estimation2
Adversarial Attack1
Atari Games1
Autonomous Driving1
Camera Pose Estimation1
Denoising1
Edge-computing1
Image Classification1
Image Generation1
Language Modeling1
Language Modelling1
Scene Understanding1
Self-Supervised Learning1
Stochastic Optimization1
Text to Image Generation1
Text-to-Image Generation1
image-classification1

Usage over time archive 2025-07-28

Papers per year tagged with Demon: 2019 to 2025, peak 5 5 0 2019: 1 paper 2019 2020: 0 papers 2020 2021: 2 papers 2021 2022: 4 papers 2022 2023: 3 papers 2023 2024: 5 papers 2024 2025: 1 paper 2025
Papers per year the archive tags with this method, by the paper's archive date (16 dated). Bars are counts, not a trend claim.

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

Momentum Rules

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