A Planning Algorithm to Support Learning in Open-ended, Unstructured Environments
Bibliographic record
Abstract
The focus of this paper is a novel pedagogical planner that we have developed called the CFLS planner (Collaborative Filtering based on Learning Sequences). The CFLS planner has been designed for an open-ended and unstructured learning environment based on the ecological approach (EA) architecture (McCalla Journal of Interactive Media in Education, 7 , 2004 ). The EA-based learning environment represents its content as learning objects (LOs), maintains models of its learners, and keeps track of learner interactions with the LOs by attaching traces of their behaviour to the LOs they have interacted with. The CFLS planner creates pedagogical plans for a target learner by looking back at the sequence of the b (for “backward”) most recent LOs that the target learner has interacted with and finding a neighbourhood of other learners who in the past have interacted with a similar sequence of b LOs. The CFLS planner then recommends to the target learner a sequence of f (for “forward”) LOs that was the most successful sequence (in terms of learning outcomes) that had been carried out next among the neighbourhood of similar learners. We implemented and tested the CFLS planner using a very simple simulation, in which simulated learners interact with simulated learning objects. We experimented with various settings for the b and f parameters. Intriguing patterns in the relationship between b and f emerged. Further, the settings for b and f that led to the best learning outcomes (on two different measures of success) varied according to the aptitude levels of the learners. Finally, we compared the CFLS planner to two baseline planners: a simple prerequisite planner (SPP) and a planner that randomly recommended the next learning object (Random). The CFLS planner readily outperformed Random (as expected), but also, more surprisingly, with appropriate settings of b and f , it outperformed SPP even though the CFLS planner did not know about the prerequisite relationships among the LOs that the SPP was able to access. This shows promise that a CFLS planner can find niche learning paths to recommend to learners based on interaction traces left behind by the learners, without needing externally engineered metadata about the learning objects or knowing very much about the learners, perhaps even finding paths based on patterns of learning activity never considered by a human designer. This is exactly the kind of planning system needed in open-ended, unstructured learning environments.
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 | 0.001 | 0.003 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.006 | 0.001 |
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".