Rational agents: prioritized goals, goal dynamics, and agent programming languages with declarative goals
Bibliographic record
Abstract
I introduce a specification language for modeling an agent's prioritized goals and their dynamics. I use the situation calculus along with Reiter's solution to the frame problem and predicates for describing agents' knowledge as my base formalism. I further enhance this language by introducing a new sort of infinite paths. Within this language, I discuss how to systematically specify prioritized goals and how to precisely describe the effects of actions on these goals. These actions include adoption and dropping of goals and subgoals. In this framework, an agent's intentions are formally specified as the prioritized intersection of her goals. The ``prioritized'' qualifier above means that the specification must respect the priority ordering of goals when choosing between two incompatible goals. I ensure that the agent's intentions are always consistent with each other and with her knowledge. I investigate two variants with different commitment strategies. Agents specified using the ``optimizing'' agent framework always try to optimize their intentions, while those specified in the ``committed'' agent framework will stick to their intentions even if opportunities to commit to higher priority goals arise when these goals are incompatible with their current intentions. For these, I study properties of prioritized goals and goal change. I also give a definition of subgoals, and prove properties about the goal-subgoal relationship. \n \nAs an application, I develop a model for a Simple Rational Agent Programming Language (SR-APL) with declarative goals. SR-APL is based on the ``committed agent'' variant of this rich theory, and combines elements from Belief-Desire-Intention (BDI) APLs and the situation calculus based ConGolog APL. Thus SR-APL supports prioritized goals and is grounded on a formal theory of goal change. It ensures that the agent's declarative goals and adopted plans are consistent with each other and with her knowledge. In doing this, I try to bridge the gap between agent theories and practical agent programming languages by providing a model and specification of an idealized BDI agent whose behavior is closer to what a rational agent does. I show that agents programmed in SR-APL satisfy some key rationality requirements.
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 | 0.004 | 0.004 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.005 |
| Scholarly communication | 0.004 | 0.008 |
| Open science | 0.003 | 0.003 |
| Research integrity | 0.002 | 0.004 |
| Insufficient payload (model declined to judge) | 0.004 | 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".