SciLayer

Discovery

Browse articles

Preprints and accepted journal articles, grouped by research program. Catalog defined in content/articles/catalog.json.

ASRA — Foundations & Theory

Architecture-level concept papers and pedagogical primers that frame the Adaptive Scientific Reasoning program before phase-specific implementation.

published preprintv1CC-BY-4.0Concept Paper

Intelligence as Representation Discovery: Ontologies, Semantics, and the Foundations of Adaptive Scientific Reasoning

Intelligence is framed not as optimization alone but as the search for increasingly useful representations of reality. Before learning can succeed, a system must discover appropriate state spaces, action semantics, evaluation criteria, and ontologies. This concept paper argues that prompts, benchmarks, world models, and ontologies are central substrates of intelligence, and introduces ASRA as a representation-first framework that infers semantic operators from observed transitions before constructing causal world models.

representation learningontology discoveryaction semanticsworld modelsbenchmarksasrascientific reasoningmodel
published preprintv1CC-BY-4.0Concept Paper

Transition-Centric Memory and Directed Exploration: Beyond Compressive Observation Memory in ASRA

Most machine learning memory optimizes compression of observations—autoencoders, predictive coding, latent state descriptions. Scientific reasoning in unknown interactive worlds requires a different substrate: transition logs, exploration graphs, and experiment records optimized for intervention, not reconstruction. This concept paper contrasts compressive observation memory with transition-centric experiment memory, explains why causal structure emerges through interaction (Pearl), situates ASRA relative to Buchanan–Ma representation learning, and shows how ASRA Phase 3 decomposes directed exploration into novelty and usefulness under a step budget—turning episodic transitions into reusable exploration graphs rather than compressed latents.

transition-centric memoryexploration memorynoveltyusefulnessasraphase 3scientific reasoningepisodic memory
published preprintv1CC-BY-4.0Concept Paper

Architectures for Adaptive Scientific Reasoning Under Uncertainty

Scientific intelligence increasingly depends on systems that reason from interventions rather than merely fit observations. This review synthesizes conceptual foundations from model-based reinforcement learning, active inference, causal inference, information theory, and perturbation biology into a unified architecture-level view of adaptive scientific reasoning under uncertainty.

adaptive scientific reasoningworld modelscausal inferenceintervention learningactive inferenceperturbation biologyaimodel

ASRA — Phase Preprints (ARC Prize 2026)

Cumulative phase articles for the nine-layer ASRA stack: experience → observation → memory → causality → goals → planning → robustness → biology bridge → integration.

Phase 1published preprintv3CC-BY-4.0Article

Transition-Centric Experience: ASRA Phase 1 — From Unknown Actions to Empirical Reasoning

Phase 1 establishes the ASRA Experience Engine: transition logging, hash-stable state IDs, effect-based action semantics inference, uncertainty-directed exploration, dead-end taboo, Swarm orchestration, and competition-grade execution fidelity. The article presents full theory and architecture; a companion Kaggle notebook deploys asra-v0.1-phase1 for ARC Prize 2026.

asraphase 1experience enginetransition modelsaction semanticsexplorationarcexecution fidelity
Phase 2published preprintv1CC-BY-4.0Article

Object-Centric Adaptive Reasoning: ASRA Phase 2 — From Pixel Transitions to Symbolic Structure

Phase 1 established interactive intelligence from transition evidence alone; Phase 2 adds the Observation Engine—segmentation, object alignment, transform events, and rule candidates over ARC demonstration pairs, with compact object-scene hints feeding back into the interactive agent. This article presents the theory, architecture, and design principles for object-centric structure between Phase 1 logging and later memory, causality, and planning.

asraphase 2object-centric reasoningarcobservation enginerule inductionadaptive reasoningconnected components
Phase 3published preprintv1CC-BY-4.0Article

Directed Exploration and Episodic Memory: ASRA Phase 3 — From Structure to Navigation

Phase 3 extends ASRA with the Navigation and Memory Engine: exploration graphs, visitation memory, novelty versus usefulness scoring, compositional subgoals, strategy reuse, and transition replay—building on Phase 1 transitions and Phase 2 object-centric observation. The article presents theory and architecture; a companion Kaggle notebook deploys compact exploration hints for ARC Prize 2026.

asraphase 3explorationepisodic memorynavigationnoveltysubgoalsarc
Phase 4published preprintv1CC-BY-4.0Article

Causal Action Semantics: ASRA Phase 4 — From Observed Effects to Intervention Reasoning

Phase 4 adds the Semantics and Causal Inference Engine: semantic signatures, causal transition models, hypothesis confirm/refute, counterfactual queries, and epistemic uncertainty over (state, action) pairs—building on Phases 1–3. The article presents theory and architecture; a companion Kaggle notebook deploys CausalSemanticsEngine hints for ARC Prize 2026.

asraphase 4action semanticscausal inferencecounterfactualshypothesis testinguncertaintyarc
Phase 5published preprintv1CC-BY-4.0Article

Goal Inference and Hypothesis Ranking: ASRA Phase 5 — From Interventions to Scientific Objectives

Phase 5 adds the Goal Inference and Hypothesis Engine: win-condition templates, progress detection, object roles, hypothesis ranking, and discriminating experiment planning—building on Phases 1–4. The article presents theory and architecture; a companion Kaggle notebook deploys GoalHypothesisEngine hints for ARC Prize 2026.

asraphase 5goal inferencehypothesis rankingwin conditionsexperiment designobject rolesprogress detection
Phase 6published preprintv1CC-BY-4.0Article

Planning and Strategy Invention: ASRA Phase 6 — From Goals to Action Sequences

Phase 6 adds the Planning & Strategy Invention layer: BFS/A* and MCTS-lite over observed transitions, a strategy library mapping goal templates to operator sequences, meta explore-exploit control, and plan repair—building on Phases 1–5. The article presents theory and architecture; a companion Kaggle notebook deploys PlanningEngine hints for ARC Prize 2026.

asraphase 6planningstrategy librarybfsmctsmeta-controllermilestone 2
Phase 7published preprintv1CC-BY-4.0Article

Robustness and Generalization: ASRA Phase 7 — From Capability to Reliability

Phase 7 adds the Robustness & Generalization layer: failure analysis, Procgen/DMLab generalization benchmarks, memory mismatch and stuck detection, action waste reduction, and an evaluation dashboard—wrapping Phase 6 planning with self-monitoring. The article presents theory and architecture; a companion Kaggle notebook deploys RobustnessEngine guards for ARC Prize 2026.

asraphase 7robustnessgeneralizationprocgendmlabstuck detectionfailure analysis
Phase 8published preprintv1CC-BY-4.0Article

Decision Biology Bridge: ASRA Phase 8 — From Grid Worlds to Perturbation–Response Reasoning

Phase 8 extends ASRA's transition-centric cognitive stack from grid worlds to perturbation–response biology: cell-state hashing, perturbation semantics, pathway hypotheses, and a biological transition graph on LINCS, OmniPath, scPerturb, and HCA context. The article presents theory and architecture; a companion Kaggle notebook deploys the asra-v0.9-phase8 bridge agent for ARC Prize 2026.

asraphase 8decision biologylincsomnipathperturbationpathway hypothesistransition graph
Phase 9published preprintv1CC-BY-4.0Article

Final Submission and Research Story: ASRA Phase 9 — Integrating Adaptive Scientific Reasoning

Phase 9 integrates Phases 1–8 into asra-v1.0-phase9: the final ARC Prize 2026 competition agent, unified architecture narrative, evaluation report framing, and Decision Biology extension section. The article presents integration theory and communication deliverables; a companion Kaggle notebook submits the v1.0 agent.

asraphase 9integrationarc prize 2026final submissionresearch storyevaluation reportdecision biology

ASRA — Technical Specifications

Implementation-oriented specs that complement phase concept papers with module boundaries, eval contracts, and deployment notes.

published preprintv1CC-BY-4.0Article

ARC-AGI-3 Kaggle Gateway Deployment Specification

Technical specification for the official ARC Prize 2026 Kaggle evaluation contract: validation vs scoring rerun, gateway sidecar integration, agent registration, submission.parquet schema, template-agent requirements, and shared ASRA notebook tooling. Documents the root cause of generic Kaggle Error failures and the gateway pattern that produced the first successful submission.

arc-agi-3arc prize 2026kagglegateway sidecardeploymentsubmission.parquetasraevaluation contract
published preprintv1CC-BY-4.0Article

Repeated-Run Learning Evaluation for ARC-AGI-3 — Protocol and ASRA Mapping

Evaluation protocol for measuring genuine learning across repeated ARC-AGI-3 episodes: full-game and single-level setups, primary metrics (action count, dead-end rate, visit redundancy), cross-run persistence requirements, and mapping to ASRA Phase 1–3 transition logging, semantics inference, and episodic memory.

arc-agi-3repeated-run learningevaluation protocoltransition modelsaction semanticsepisodic memoryasraadaptive reasoning

ASRA — Evaluation & Results

Empirical evaluation reports, benchmark results, and competition score consolidations.

published preprintv1CC-BY-4.0Concept Paper

Adaptive Scientific Discovery Benchmark (ASDB): A Two-Track Framework for Evaluating Interactive Agents

Most agent benchmarks assume documented tool semantics and static ground-truth answers. Real scientific inquiry requires agents to learn what interventions do from state transitions, then infer hidden mechanisms, design discriminating experiments, and predict held-out observables. ASDB unifies two complementary tracks: Action Semantics Discovery (inferring an action map φ̂(a) from unlabeled controls) and Scientific Discovery Evaluation (recovering hidden theory classes under an intervention budget). Both share one interaction loop but score different constructs. Linked A→B episodes, decoy falsification, tiered difficulty, and decomposable metrics aim at construct validity for adaptive scientific reasoning evaluation.

asdbscientific discovery benchmarkaction semantics discoveryconstruct validityinteractive agentsintervention designinverse problemsadaptive scientific reasoning

Comparative & Position Papers

External framework comparisons and synthesis essays that situate ASRA relative to adjacent research programs.

published preprintv1CC-BY-4.0Concept Paper

ASRA vs Buchanan–Ma: A Mathematical Theory of Memory

Comparative analysis of Buchanan, Pai, Wang, and Ma's representation-learning textbook (arXiv:2606.06624) and the Adaptive Scientific Reasoning Architecture (ASRA). Buchanan–Ma formalizes compressive memory and white-box deep representations; ASRA formalizes intervention loops, causal semantics, goal hypotheses, and experiment design. The programs are complementary: memory theory vs scientific reasoning under uncertainty.

asrarepresentation learningmathematical theory of memoryworld modelscompressioncausal inferencescientific intelligencebuchanan

Nature Foundation Models

Organizational and scientific framing for Nature Foundation Models research themes.

published preprintv2CC-BY-4.0Concept Paper

Nature Foundation Models: A Hierarchical Framework for Learning Worlds, Embodiment, and Scientific Intelligence

We propose Nature Foundation Models (NFM), a research program for systems that learn representations, dynamics, causal structure, and mechanisms directly from interaction with the natural world. NFM organizes scientific intelligence as a hierarchy—NFM-Worlds, NFM-Robotics, Atlas, and Atlas-GS—with a shared state–action–dynamics abstraction and a seven-stage developmental pipeline from world representation to adaptive scientific reasoning. The central thesis is that scientific reasoning should emerge from increasingly sophisticated interactions with learned world models rather than from an independent symbolic module.

nature foundation modelsworld modelsembodied aiscientific reasoningcausal discoverynfm-roboticsatlas-gsgaussian splatting
published preprintv1CC-BY-4.0Implementation Paper

Atlas-GS: An End-to-End Implementation of Gaussian World Modeling for Embodied Robotics

We present Atlas-GS v1, the first end-to-end implementation of a 3D Gaussian world-modeling pipeline within the Nature Foundation Models (NFM) hierarchy. Atlas-GS ingests RGB-D observations, constructs a persistent Gaussian field, localizes against that map, persists world state across sessions, and logs state–action–state transitions. The system implements Phases 0–6 as a modular Python package with CLI tools and demo video generation—without requiring GPU hardware for v1 validation. We report empirical results on TUM RGB-D (4,018 Gaussians, 0.0102 m localization RMSE) and synthetic sequences.

atlas-gs3d gaussian splattingworld modelsembodied airgb-d mappingscene memorynature foundation modelsrobotics
published preprintv1CC-BY-4.0Implementation Paper

Orbit Wars Phase 4: Applying NFM, ASRA, and Atlas-GS to Multi-Agent RTS Competition

We present Orbit Wars Phase 4, an open-source Kaggle agent unifying Nature Foundation Models (state–action–dynamics world modeling), Atlas-GS (2D Gaussian spatial value splat for target prioritization), and ASRA (hypothesis-driven simulation with five strategic theories tested per turn). The agent operates under a one-second decision budget in a continuous 2D RTS with orbiting planets, comets, and multi-agent free-for-all play.

orbit warsnature foundation modelsasraatlas-gsworld modelsgame aikagglemulti-agent rts

Platform & Industry Essays

Essays on ML platform evolution, tooling, and engineering practice outside the core ASRA arc.

published preprintv2CC-BY-4.0Concept Paper

From "Databricks Meets PyTorch" to "Databricks Meets Keras": AI Platform Evolution with ASRA

A practical technical essay on two recurring AI platform patterns—enterprise data infrastructure paired with maximum research flexibility (PyTorch-class) versus the same substrate with radical simplicity (Keras-class)—and how ASRA implements both lanes on the path toward scientific intelligence platforms.

ai infrastructuremlopspytorchkeraslakehousescientific aiasradecision biology

Other articles