Open research questions in AI-based Problem Solving and Planning

What the literature leaves open

• The need to consider the relative weights of risk factors in risk analysis. The challenge of selecting representative samples for expert labeling.

  A data-driven cost-sensitive sorting method with active learning for ordinal risk classification of failure modes · 2026 · DOI

• The need for a data-driven and cost-sensitive approach to risk classification. The lack of a method that combines clustering trees and semi-supervised cost-sensitive decision trees for FMEA.

  A data-driven cost-sensitive sorting method with active learning for ordinal risk classification of failure modes · 2026 · DOI

• The lack of a unified assessment and analytical framework for disease pathway analysis in aquaculture. The limited systematic quantitative analyses of disease costs and benefits of control measures. The need for a framework that can perform zero-shot causal reasoning for aquaculture diseases.

  DPAIA: a zero-shot multi-agent causal reasoning framework for interpretable disease pathway analysis in aquaculture · 2026 · DOI

• The high_risk_unresolved reference contains only one positive story, which is insufficient for evaluating high-risk detection capability; the field is retained as an operational part of the framework but not used to support a detection claim.

  Evidence-Aware Human-in-the-Loop LLM Review for Requirements-to-Planning Decisions · 2026 · DOI

• Future work will focus on addressing these limitations, with a primary emphasis on transitioning the framework from simulation to real-world hardware. Third, although its effectiveness was validated in 4v4 and 2v2 scenarios, its generalization capability in larger-scale engagements against more complex competitive strategies remains to be further explored.

  CaST-MASAC: Integrating Causal Inference and Spatio-Temporal Attention for Multi-UAV Cooperative Task Planning · 2025 · DOI

• Whether agency ultimately joins procedures, objects, processes, and services in the canon of durable organizing abstractions remains to be seen.

  Agency as a New Organizing Abstraction for Computer Science · 2026 · DOI

• The primary gap addressed by this work is the lack of fully functional, domain-specific applications of agentic AI. Prior research has addressed autonomous reasoning frameworks, multi-agent orchestration strategies, and safety mechanisms, but few studies have implemented these principles in consumer-facing applications.

  A multi model orchestrated agent for live flight and hotel itinerary generation · 2026 · DOI

• The paper identifies several limitations, including the fact that the dataset is intentionally weakly supervised and that the model is trained on a single workstation GPU. The authors note that student-scale research is often compute-limited rather than idea-limited.

  Scaling Laws and Architectural Advances of Hierarchical JEPA (H-JEPA) Model for Planning, Control and Robotics in Physical Systems · 2026 · DOI

• The paper suggests several future research directions, including incorporating language-conditioned abstract sub-goals into H-JEPA and incorporating uncertainty into the planning process. The authors suggest that H-JEPA can be used as a bridge between latent planning and instruction following.

  Scaling Laws and Architectural Advances of Hierarchical JEPA (H-JEPA) Model for Planning, Control and Robotics in Physical Systems · 2026 · DOI

• The need for transparent and interpretable decision-making in MAPF. The lack of a unified semantic schema for MAPF.

  maPO: An Ontology for Multi-Agent Path Finding and Its Usage for Explaining Planner Behaviour · 2026 · DOI

• This model has several limitations. First, ROM‑level equivalence is not full semantic equivalence. It is equivalence under a declared canonical‑ ization contract. Natural‑language meaning remains broader than any deterministic canonical form. Second, canonicalization intentionally discards non‑semantic representation entropy.

  AIKernel Phase-1 Paper 01: ROM Format and Knowledge Snapshot Model · 2026 · DOI

• The lack of constitutional governance infrastructure in long-term AI-assisted operational systems. The need for topology-aware operational cognition in intelligent systems.

  Persistent Operational Memory Through Knowledge Engine Architecture · 2026 · DOI

• Crucially, even when reflection substantially increases the number of fixed_by edges, these relations remain sparse relative to overall graph size, indicating that EXG-based methods rely on amplifying a limited number of well-aligned, high-impact fixes rather than on frequent trial-and-error corrections.

  EXG: Self-Evolving Agents with Experience Graphs · 2026

• Current scaling practices focus on collecting more experience or broader task sets under fixed benchmarks. There is a lack of systematic exposure to environments with meaningfully different executable rule-sets. Generalizable agents require environment scaling to adapt to diverse tasks and unseen environments.

  Scalable Environments Drive Generalizable Agents · 2026

• The human preference study is constrained by time and budget, resulting in a small number of annotators and a limited sample size. Applying MA$^{2}$P to a new domain requires a warm-up phase. Most of the automatic metrics rely on an LLM evaluator.

  MA$^{2}$P: A Meta-Cognitive Autonomous Intelligent Agents Framework for Complex Persuasion · 2026

• 024 strategic calls per step under event-driven triggers, but causes the largest success drop, showing that event-driven triggers alone are insufficient for long-horizon recovery.

  SPIKE: An Adaptive Dual Controller Framework for Cost-Efficient Long-Horizon Game Agents · 2026

• Limited dependence on operations research experts. Need for interactive and continuous adaptation of deployed optimization models.

  Democratizing Large-Scale Re-Optimization with LLM-Guided Model Patches · 2026

• While we have isolated the validator’s contribution via a controlled experiment (loop active, validator disabled; §5.3), controlled comparisons of lazy vs. full schema injection across model sizes—particularly on SGD where the 26-service schema amplifies context load—and always-on history inclusion vs. on-demand retrieval via τH remain as future work.

  ReacTOD: Bounded Neuro-Symbolic Agentic NLU for Zero-Shot Dialogue State Tracking · 2026

• The successful completion of subtasks depends critically on the tight coordination between the Action Corrector and the Subtask Actor, ensuring that corrective adjustments and action generation are effectively aligned.

  HiRBP: a VLM-based hierarchical decision-making framework for long-horizon robotic behavior planning (ChinaMM 2026) · 2026 · DOI

• To develop more advanced meta-cognitive configurators that can select more effective meta-strategies. To explore the use of MA$^{2}$P in other applications, such as education and behavior change.

  MA$^{2}$P: A Meta-Cognitive Autonomous Intelligent Agents Framework for Complex Persuasion · 2026

• Network infrastructure operation is complex and challenging. LLMs are probabilistic models that require clarification of their capabilities. Complex tasks remain difficult for LLMs without additional reasoning.

  How Helpful is LLM Assistance in Network Operations? A Case Study at a Large Demonstration Network · 2026

• The Ticket function was used less frequently due to its limited design. Complex tasks remain difficult for the LLM when used without additional reasoning.

  How Helpful is LLM Assistance in Network Operations? A Case Study at a Large Demonstration Network · 2026

• Validating the dataset using large-scale real-time data collected from operational aquaculture environments - Improving the flexibility of the framework - Conducting evaluations in real-time…

  DPAIA: a zero-shot multi-agent causal reasoning framework for interpretable disease pathway analysis in aquaculture · 2026 · DOI

• Investigating how to generate interaction data that reliably drives continual improvement. Exploring the use of scalable environments as a training substrate for general agents. Developing new methods for environment scaling and stateful learning mechanisms.

  Scalable Environments Drive Generalizable Agents · 2026

• Existing methods struggle to convert prior interactions into portable decision strategies - Embodied agents need to integrate visual perception, semantic goals, spatial context, and action consequences under partial…

  Robo-Cortex: A Self-Evolving Embodied Agent via Dual-Grain Cognitive Memory and Autonomous Knowledge Induction · 2026

Most-cited papers in AI-based Problem Solving and Planning

• Code as Policies: Language Model Programs for Embodied Control · 2023 · 638 citations
• ProgPrompt: Generating Situated Robot Task Plans using Large Language Models · 2023 · 510 citations
• Using cognitive psychology to understand GPT-3 · Proceedings of the National Academy of Sciences · 2023 · 443 citations
• Artificial cognition for social human–robot interaction: An implementation · Artificial Intelligence · 2016 · 316 citations
• Computational planning of the synthesis of complex natural products · Nature · 2020 · 254 citations
• Accelerating Innovation With Generative AI: AI-Augmented Digital Prototyping and Innovation Methods · IEEE Engineering Management Review · 2023 · 252 citations
• The increasing cost tree search for optimal multi-agent pathfinding · Artificial Intelligence · 2012 · 195 citations
• Towards a science of integrated AI and Robotics · Artificial Intelligence · 2017 · 145 citations
• If you don't want to be late, enumerate: Unpacking reduces the planning fallacy · Journal of Experimental Social Psychology · 2003 · 139 citations
• Representations for robot knowledge in the KnowRob framework · Artificial Intelligence · 2015 · 129 citations

Most recent work

• Mirror Observerhood Lab VI: Neuro-Symbolic Persistence in Language-Like Agents · Zenodo (CERN European Organization for Nuclear Research) · 2026
• A Comprehensive Survey of Agents for Computer Use: Foundations, Challenges, and Future Directions · Journal of Artificial Intelligence Research · 2026
• The Lever Generalizes -- and It Brakes: A Late, Bidirectional Action-Commitment Lever Across Agent Decisions and Architectures · Zenodo (CERN European Organization for Nuclear Research) · 2026
• Towards generalizable fault diagnosis via LLM-driven hierarchical cross-modal alignment · Knowledge-Based Systems · 2026
• EvoMind: A Governed Cognitive Architecture for Persistent, Verifiable, Experience-Driven Agentic Autonomy · Zenodo (CERN European Organization for Nuclear Research) · 2026
• Long causal chain-of-thought for compound fault diagnosis: enhancing hypergraph-based Causal-LLM intervention reasoning · Journal of Manufacturing Systems · 2026
• A domain knowledge-enhanced MBSE framework for developing knowledge-based inspection systems: A case study on overhead crane inspection · Computers & Industrial Engineering · 2026
• Case-Based Reasoning Augmented Large Language Model Framework for Decision Making in Realistic Safety-Critical Driving Scenarios · Safety Science · 2026
• A Hydro-Intelligence Body: Cognitive Architecture for Project-Oriented Autonomous Water Systems Driven by Multi-Modal Sensing and Water Utility Panel · IEEE Transactions on Consumer Electronics · 2026
• A fault diagnosis model based on belief rule base with dynamic uncertainty · Information Sciences · 2026

Related topics in Computer Science