ASE+ACE

Closed weights University of Massachusetts Amherst 0.3K parameters September 1983

No estimate

No hardware requirements for this model

The weights for this model have not been published, so it cannot be downloaded or run on your own hardware at any size. It is reachable only through its provider, and no graphics card changes that.

On record

Full specification

Everything on record for this model. Most of it describes how it was trained rather than how it runs — useful context for judging how much work went into it, and how it compares with models built at a different scale.

Origin

Who built this model, where, and when it was published.

Organisation
University of Massachusetts Amherst
Organisation type
Academia
Country
United States of America
Published
1 September 1983
Authors
Andrew G. Barto, Richard S. Sutton, and Charles W. Anderson

What it does

The problem areas the model was built for. A model can carry several of each.

Domain
Robotics
Task
Pole balancing

Size

How large the model is and how much data it was trained on. Parameters are the figure that decides whether it fits on a given graphics card.

Parameters
0.3K

The system consists of two parts: ACE and ASE, each with 162 weights (=324 parameters). Found in Figures 2 and 3.

Training data
500,000 tokens

"Runs consisted of 100 trials unless the run's duration exceeded 500 000 time steps (approximately 2.8 h of simulated real time)" "Almost all runs of the ASE/ACE system [...], were terminated after 500 000" (Section IX)

Training compute

The arithmetic performed to train the model, measured in floating-point operations. It is a measure of what the training run cost, not of how fast the finished model answers you.

Training compute
3.2 × 10⁸ FLOP

324 * 2 * 500000 = 324000000 = 3.24e8. The calculation assumes "compute per forward pass" = "number of parameters" = "compute per backward pass". Their model only has a single layer and is trained with simple update rules instead of gradient descent. Training details are described in Section IX. Note that this is the compute for a single run; they appear to have repeated training 10 times for the ASE+ACE system.

How it was established
Operation counting

The training run

What it physically took to train: which chips, how many, for how long, and what that drew from the wall.

Wall-clock time
3 hours

"Runs consisted of 100 trials unless the run's duration exceeded 500 000 time steps (approximately 2.8 h of simulated real time)" "Almost all runs of the ASE/ACE system [...], were terminated after 500 000" (Section IX)

How it is classified

Labels the source dataset applies when tracking notable models, and how confident it is in the entry.

Frontier model
Yes
Why it is tracked
Highly cited,Historical significance
Record confidence
Likely
Citations
4,296

Sources

Where this record came from and when it was last checked.

Reference
Neuronlike adaptive elements that can solve difficult learning control problems
Last updated
28 November 2025

What the numbers mean

About this model

ASE+ACE was published by University of Massachusetts Amherst, in the country recorded as United States of America, during September 1983. The category the publisher falls under is academia.

It works in the domain of Robotics, and is recorded as performing the task of pole balancing.

This is a closed model: the trained values stayed with whoever produced them, and there is no local version to run.

What went into building it

Training it took a computation budget of roughly 3.2 × 10⁸ FLOP. That figure measures what producing the model cost, and has no bearing on how fast it answers.

It was trained on a corpus of about 500,000 tokens of text.

The reason it appears in this catalogue at all: highly cited,Historical significance.

Answers

ASE+ACE — common questions

01

ASE+ACE— how much compute was used to train it?

Training consumed around 3.2 × 10⁸ FLOP. That measures what producing the model cost and says nothing about how quickly it answers once trained — inference speed comes from memory bandwidth, not from the training budget.

02

ASE+ACE— what GPU do I need to run it?

None. This is a closed model — its weights were never published, so it cannot be downloaded or run on your own hardware at any price. It is reachable only through its provider.

03

ASE+ACE— is it open source?

The licensing was never recorded in our source data. We treat unstated licensing as closed, because an unrecorded licence is not one to rely on.

04

ASE+ACE— how many parameters does it have?

It has a parameter count of 0.3K. The system consists of two parts: ACE and ASE, each with 162 weights (=324 parameters). Found in Figures 2 and 3. That figure is the total, and it is what decides how much memory the model needs — roughly half a gigabyte per billion at the compression most people use.

05

ASE+ACE— who created it?

It was published by University of Massachusetts Amherst, based in United States of America, an organisation categorised as academia.

06

ASE+ACE— when was it released?

It was published in September 1983. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.

07

ASE+ACE— what is it used for?

It works in the domain of Robotics, and is recorded as handling the task of pole balancing. A model can carry several of each, so these are the areas it was built for rather than a limit on what it will attempt.

Source

Original publication

Record last updated 28 November 2025

The other direction

Looking at it from the other side?

This page starts from the model. If you already own a card and want to know everything it will run, start from the hardware instead.