Data collection using deep reinforcement learning for serious games
Bibliographic record
Abstract
Mild Cognitive Impairment (MCI) often occurs prior to the more serious condition of dementia and early detection of MCI is an important but challenging task because of its indistinct symptoms. Work has been done developing serious games on mobile devices for MCI detection as opposed to a typical application of serious games for growing and maintaining mental acuity. The serious games WarCAT and Locker record player’s moves made while playing the game to determine their levels of strategy recognition and learning. To be able to demonstrate this, however, requires a large amount of player data. Therefore, it would be beneficial to develop a method of generating synthetic data that could imitate human player data. The area of machine learning (ML) known as Reinforcement Learning (RL) can be applied to creating a large pool of players since it emulates the way humans learn. In RL, if an action in response to a stimulus is followed by a successful reward, the stimulus-action-reward association will be strengthened, and the reward will be recalled with greater likelihood upon later presentation of the same stimulus and action. Like RL in humans, considerable trial and error (training) is often required. In addition, a growing subfield of machine learning known as Deep Reinforcement Learning uses techniques of Reinforcement Learning along with Artificial Neural Networks for function approximation. The purpose of this thesis is to explore the use of Deep RL to learn to play our serious game and achieve gameplay results comparable to the best human performance. From this, we can define baselines which allow us to create bots with various levels of training to emulate individuals at various stages of learning, or by extension, various levels of cognitive decline.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.014 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.004 | 0.003 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".