arXiv research

Improves model accuracy for neural nets in stochastic dynamics with partial prior knowledge.

Arriving at the complete probabilistic knowledge of a domain, i.e., learning how all variables interact, is indeed a demanding task. In reality, settings often arise for which an individual merely possesses partial knowledge of the domain, and yet, is expected to give adequate answers to a variety of posed queries. Tha…

New method learns from non-uniform data and partial physical knowledge.

RID framework quantifies and regularizes task-relevant knowledge in distillation.

Reinforcement learning would enjoy better success on real-world problems if domain knowledge could be imparted to the algorithm by the modelers. Most problems have both hidden state and unknown dynamics. Partially observable Markov decision processes (POMDPs) allow for the modeling of both. Unfortunately, they do not p…

New method learns frequency-dependent partial correlations.

Novel framework for learning infinitesimal generator of stochastic processes.

Reconstructing a planar domain from its Dirichlet-to-Neumann data

Enhancing spectral embedding for low-dimensional embeddings in rare disease cohorts

Proposes SPCA to incorporate structural constraints in model identification.

Cluster-DAGs improve causal discovery with prior knowledge.

Many interesting real world domains involve reinforcement learning (RL) in partially observable environments. Efficient learning in such domains is important, but existing sample complexity bounds for partially observable RL are at least exponential in the episode length. We give, to our knowledge, the first partially …

Recent work in learning ontologies (hierarchical and partially-ordered structures) has leveraged the intrinsic geometry of spaces of learned representations to make predictions that automatically obey complex structural constraints. We explore two extensions of one such model, the order-embedding model for hierarchical…

Centuries of development in natural sciences and mathematical modeling provide valuable domain expert knowledge that has yet to be explored for the development of machine learning models. When modeling complex physical systems, both domain knowledge and data provide necessary information about the system. In this paper…

Oracle-efficient algorithm for offline RL with partial data coverage.

We prove that if $(M,g)$ is a topological 3-ball with a $C^4$-smooth Riemannian metric $g$, and mean-convex boundary $\partial M$ then knowledge of least areas circumscribed by simple closed curves $γ\subset \partial M$ uniquely determines the metric $g$, under some additional geometric assumptions. These are that $g$ …

KalmanNet uses neural networks to improve state estimation in systems with unknown dynamics.

Improved CI test for heteroskedastic data enhances causal discovery.

Proposes a method to identify causal relationships using background knowledge.

We infer both microscopic and macroscopic behaviors of a three-dimensional chaotic fluid flow using reservoir computing. In our procedure of the inference, we assume no prior knowledge of a physical process of a fluid flow except that its behavior is complex but deterministic. We present two ways of inference of the co…

Introduces a rule-based Bayesian regression for better uncertainty quantification and expert knowledge integration.

New bounds for PDA using partial optimal transport improve domain alignment.

Enhances neural operators with physics knowledge for more accurate simulations.

The paper develops methods to estimate POMDPs from partial information.

New method for comparing different mass measures on tree structures using entropy partial transport.

b-LOAD extends local causal discovery with prior knowledge, improving causal effect estimation.

Since time immemorial, people have been looking for ways to organize scientific knowledge into some systems to facilitate search and discovery of new ideas. The problem was partially solved in the pre-Internet era using library classifications, but nowadays it is nearly impossible to classify all scientific and popular…

Integrates skills and world models for efficient task solving and transfer.

Framework uses physics knowledge to improve spatiotemporal prediction with limited data.

New algorithms solve partial optimal transport problems for applications like PU learning.

New algorithm proves RL from partial obs is feasible.

The paper analyzes extreme risk measures with limited distributional information.

A tutorial on optimizing complex functions with partial knowledge.

Extracts geometric information from point-clouds for multiclass classification.

Transformer learns to infer partial MDPs for efficient in-context adaptation and exploration.

Identifies causal effects in partially directed acyclic graphs with observed variables.

ParKCa combines multiple causal inference methods to infer new causes from known and unknown factors.

Paper tackles robust knowledge transfer in parallel RL tasks.

One goal in Bayesian machine learning is to encode prior knowledge into prior distributions, to model data efficiently. We consider prior knowledge from systems of linear partial differential equations together with their boundary conditions. We construct multi-output Gaussian process priors with realizations in the so…

The paper develops algorithms for competitive RL in partially observable MGs.

We derive upper bounds on the complexity of ReLU neural networks approximating the solution maps of parametric partial differential equations. In particular, without any knowledge of its concrete shape, we use the inherent low-dimensionality of the solution manifold to obtain approximation rates which are significantly…

This paper introduces glocal explanations for expected goal models in soccer.

Paper characterizes and represents pairwise causal background knowledge for improved causal inference.

Optimistic Q-learning reduces sample complexity for systems with known dynamics.

This work improves knowledge distillation by transferring full kernel matrices efficiently.

Derives Black-Scholes model without stochastic calculus or PDEs.

Data-driven models analyze power grids under incomplete physical information, and their accuracy has been mostly validated empirically using certain training and testing datasets. This paper explores error bounds for data-driven models under all possible training and testing scenarios, and proposes an evaluation implem…

New method achieves optimal performance without needing problem parameters.