Flat minima are illusory; generalization is driven by weakness, a reparameterization-invariant measure of compatible completions that predicts performance better than sharpness on MNIST and Fashion-MNIST.
hub
A formal basis for the heuristic determination of minimum cost paths
19 Pith papers cite this work. Polarity classification is still indexing.
19
Pith papers citing it
hub tools
citation-role summary
background 2
method 1
citation-polarity summary
representative citing papers
Tree of Thoughts enables language models to solve complex planning tasks by generating, evaluating, and searching over coherent intermediate thoughts in a tree, raising Game of 24 success from 4% to 74% with GPT-4.
cGAN surrogates recover 45-60% of CFD energy savings and high-velocity wake avoidance in 3D AUV path planning while running at 28-146 microsecond inference speeds across 19,800 trajectories.
OSCAR learns class-conditioned survival distributions for obstacle clearance times online (handling right-censored data) to compute patience thresholds in graph-based navigation, converging near oracle performance after few observations per class.
Zeta* and Zeta*-SIPP use elliptical expansion and improved visibility to deliver optimal any-angle paths on grids, with Zeta*-SIPP over 20x faster than prior dynamic planners.
The anti-lexicographic SUS-anchor achieves sampling densities less than 1% above the lower bound for alphabet size 4 and k=1, substantially outperforming bidirectional anchors.
Bidirectional Evolutionary Search augments autoregressive expansion with evolutionary recombination operators and dense backward subgoal feedback to produce better candidates than standard best-of-N or tree search for language model self-improvement.
GRAFT-ATHENA projects combinatorial method choices into factored trees that embed as fingerprints in a metric space, enabling an agentic system to accumulate experience across domains and autonomously discover new numerical techniques for physics-informed problems.
CogInstrument represents human reasoning as revisable cognitive motifs in graphical form to support iterative alignment with LLMs during planning tasks, with a N=12 study indicating gains in targeted revision, agency, and trust over standard dialogue interfaces.
A branch-and-bound algorithm with custom node selection, branching rules, and conflict definitions solves the logic-constrained shortest path problem for flight planning with traffic flow restrictions, showing order-of-magnitude speedups on a public global dataset with 20000 real constraints.
The study introduces SSPPV and SSPPE variants of solution-space path planning for en-route ATC, integrating distance-, time-, and zone-based conflict detection, with SSPPV plus zone detection achieving 3.69 ms average computation in MUAC Delta sector simulations on a 5 nmi grid.
IR-SIM is a YAML-defined simulator for mobile robot navigation that supports text-prompt scenario creation, policy training, benchmarking, and bridging to higher-fidelity or real-world settings.
Presents a framework for training empirically admissible neural heuristics via underestimating Bellman operator, asymmetric loss, and validation calibration offset, reporting reduced node expansions with no observed admissibility violations on small puzzles.
A hybrid LLM-plus-physics-simulation framework generates synthesis routes for niobium oxides and finds that LLM implicit priors produce more viable plans than classical path-planning algorithms in computational tests.
Convex-Neural RRT* uses neural guidance to predict waypoint regions, extracts convex sampling areas from them, and reports 30-75% faster planning than other neural RRT* methods with ~5% shorter paths and >99% success rate on 18 maps.
Distill refines user task specifications for robots by pruning unnecessary steps, generalizing meanings, and relaxing order constraints, as demonstrated in a crowdsourcing study on a web interface.
A hybrid PLL oracle and neural A* method constrains LLM generation to paths in a 700K-node medical graph, claiming better latency-recall tradeoffs and fewer hallucinations than text-only RAG on fertility queries.
Genetic algorithm with AHP produces more diverse paths than A* whose runtime is independent of environment size, claimed to increase continuum robot resilience in two simulated settings.
citing papers explorer
-
Are Flat Minima an Illusion?
Flat minima are illusory; generalization is driven by weakness, a reparameterization-invariant measure of compatible completions that predicts performance better than sharpness on MNIST and Fashion-MNIST.
-
Tree of Thoughts: Deliberate Problem Solving with Large Language Models
Tree of Thoughts enables language models to solve complex planning tasks by generating, evaluating, and searching over coherent intermediate thoughts in a tree, raising Game of 24 success from 4% to 74% with GPT-4.
-
3D Underwater Path Planning via Generative Flow Field Surrogates
cGAN surrogates recover 45-60% of CFD energy savings and high-velocity wake avoidance in 3D AUV path planning while running at 28-146 microsecond inference speeds across 19,800 trajectories.
-
OSCAR: Obstacle Survival Curves for Adaptive Robot Navigation
OSCAR learns class-conditioned survival distributions for obstacle clearance times online (handling right-censored data) to compute patience thresholds in graph-based navigation, converging near oracle performance after few observations per class.
-
Optimal any-angle path planning in static and dynamic environments
Zeta* and Zeta*-SIPP use elliptical expansion and improved visibility to deliver optimal any-angle paths on grids, with Zeta*-SIPP over 20x faster than prior dynamic planners.
-
The anti-lexicographic SUS-anchor: a near-optimal k=1 sampling scheme
The anti-lexicographic SUS-anchor achieves sampling densities less than 1% above the lower bound for alphabet size 4 and k=1, substantially outperforming bidirectional anchors.
-
Self-Improving Language Models with Bidirectional Evolutionary Search
Bidirectional Evolutionary Search augments autoregressive expansion with evolutionary recombination operators and dense backward subgoal feedback to produce better candidates than standard best-of-N or tree search for language model self-improvement.
-
GRAFT-ATHENA: Self-Improving Agentic Teams for Autonomous Discovery and Evolutionary Numerical Algorithms
GRAFT-ATHENA projects combinatorial method choices into factored trees that embed as fingerprints in a metric space, enabling an agentic system to accumulate experience across domains and autonomously discover new numerical techniques for physics-informed problems.
-
CogInstrument: Modeling Cognitive Processes for Bidirectional Human-LLM Alignment in Planning Tasks
CogInstrument represents human reasoning as revisable cognitive motifs in graphical form to support iterative alignment with LLMs during planning tasks, with a N=12 study indicating gains in targeted revision, agency, and trust over standard dialogue interfaces.
-
Logic-Constrained Shortest Paths for Flight Planning
A branch-and-bound algorithm with custom node selection, branching rules, and conflict definitions solves the logic-constrained shortest path problem for flight planning with traffic flow restrictions, showing order-of-magnitude speedups on a public global dataset with 20000 real constraints.
-
Solution space path planning for supporting en-route air traffic control
The study introduces SSPPV and SSPPE variants of solution-space path planning for en-route ATC, integrating distance-, time-, and zone-based conflict detection, with SSPPV plus zone detection achieving 3.69 ms average computation in MUAC Delta sector simulations on a 5 nmi grid.
-
IR-SIM: A Lightweight Skill-Native Simulator for Navigation, Learning, and Benchmarking
IR-SIM is a YAML-defined simulator for mobile robot navigation that supports text-prompt scenario creation, policy training, benchmarking, and bridging to higher-fidelity or real-world settings.
-
Learning Empirically Admissible Neural Heuristics for Combinatorial Search
Presents a framework for training empirically admissible neural heuristics via underestimating Bellman operator, asymmetric loss, and validation calibration offset, reporting reduced node expansions with no observed admissibility violations on small puzzles.
-
Coupling Language Models with Physics-based Simulation for Synthesis of Inorganic Materials
A hybrid LLM-plus-physics-simulation framework generates synthesis routes for niobium oxides and finds that LLM implicit priors produce more viable plans than classical path-planning algorithms in computational tests.
-
Convex-Neural RRT*: Fast and Reliable Learning-Guided Sampling for High-Quality Robot Path Planning
Convex-Neural RRT* uses neural guidance to predict waypoint regions, extracts convex sampling areas from them, and reports 30-75% faster planning than other neural RRT* methods with ~5% shorter paths and >99% success rate on 18 maps.
-
Distill: Uncovering the True Intent behind Human-Robot Communication
Distill refines user task specifications for robots by pruning unnecessary steps, generalizing meanings, and relaxing order constraints, as demonstrated in a crowdsourcing study on a web interface.
-
TTFT-Aware Graph Chain-of-Thought:Distance-Indexed Neural A* for Low-Hallucination Multi-Hop Medical Reasoning
A hybrid PLL oracle and neural A* method constrains LLM generation to paths in a 700K-node medical graph, claiming better latency-recall tradeoffs and fewer hallucinations than text-only RAG on fertility queries.
-
Increasing Resilience of Continuum Robots via Motion Planning Algorithms
Genetic algorithm with AHP produces more diverse paths than A* whose runtime is independent of environment size, claimed to increase continuum robot resilience in two simulated settings.
- Human Ancestries Simulation and Inference: a Review of Ancestral Recombination Graph-Based Approaches