New approach to avoid bad incentives in reinforcement learning agents.
problem Designing safe reinforcement learning agents that avoid unnecessary disruptions.
method Break down side effects penalties into baseline state and deviation measure; introduce new stepwise inaction baseline and relative reachability deviation measure.
result Combination of new design choices avoids undesirable incentives, while simpler alternatives fail.
GraphReach improves GNN performance by incorporating node positions.
problem Existing GNNs fail to capture node positions, leading to inaccurate predictions.
method GraphReach uses reachability estimations from anchor nodes to capture global node positions.
result GraphReach achieves up to 40% relative improvement in accuracy compared to state-of-the-art GNNs.
Polynomial-time reachability for LTI systems with TLL NN controllers is achieved.
problem Bounding the reachable set of LTI systems controlled by TLL NN controllers.
method Polynomial-time computation of exact one-step reachable set and tight bounding box via two methods.
result Exact reachability computation in polynomial time for TLL NN controllers.
New algorithm verifies deep neural networks with guaranteed bounds.
problem Verifying deep neural networks' correctness is hard.
method Adaptive nested optimisation for reachability analysis.
result Efficiently verifies a broader class of DNNs than previous methods.
This paper uses normalizing flows to approximate transport maps between densities.
problem Approximating transport maps between given densities.
method Construct time-dependent controls using normalizing flows.
result Provides bounds on the number of switches for piecewise constant approximations.
Study improves sample complexity for finding influential seed nodes in networks.
problem Determining the most influential set of nodes in a network with limited simulations.
method Upper bounds on reachability variance and data-adaptive simulation method.
result Significantly improved upper bound on sample complexity for IM.
We quantify content availability and user discovery opportunities in recommender systems.
problem Determining the maximum probability of recommending content to users.
method Stochastic reachability to compute upper bounds on recommendation likelihood.
result Reachability metrics can detect biases and diagnose user discovery limitations.
Lumbermark clusters data robustly, slicing limbs of mutual reachability trees.
problem Detecting clusters of varying sizes, densities, and shapes in datasets.
method Iteratively chops limbs of a mutual reachability minimum spanning tree.
result Lumbermark produces partitions with user-specified sizes.
Guaranteed reachable set for unknown nonlinear systems on manifolds.
problem Determining reachable set for unknown nonlinear systems on manifolds.
method Underapproximations of reachable set using local dynamics and bounds on dynamics rate of change.
result Guaranteed set of reachable states for systems on complete Riemannian manifolds.
This paper tackles data-efficient nonlinear control in Hamiltonian systems using symplectic geometry.
problem Data-efficient nonlinear control in Hamiltonian systems.
method Combines symplectic geometry, recurrence on energy level sets, and chain policies to solve target reachability problems.
result Data requirements depend on geometric and recurrence properties of the Hamiltonian, not the state dimension.
We verify CNNs using reachability analysis and transformations.
problem Verifying neural-based perception systems implemented by CNNs.
method Reachability analysis for feed-forward neural networks with MILP encodings.
result The notion of local robustness cannot be captured by previous robustness notions.
Study on Heisenberg group's Lorentzian problems using Pontryagin's principle.
problem Lorentzian problems on the Heisenberg group.
method Applied Pontryagin's maximum principle to obtain extremal trajectories.
result Parameterization of abnormal and normal extremal trajectories, investigation of reachability sets and existence of optimal trajectories.
CIfly simplifies causal inference tasks with linear-time reachability primitives.
problem Efficiently solving complex causal inference problems.
method Formalizes reachability as a core operation, builds on state-space graphs, and uses rule tables.
result CIfly algorithms run in linear time, outperforming existing methods.
Procedure verifies if machine learning models assign fixed predictions that preclude access.
problem Models assign fixed predictions that preclude access to credit and employment.
method Model-agnostic recourse verification with reachable sets.
result Models can inadvertently preclude access by assigning fixed predictions.
Study analyzes and enhances robustness of neural networks for classification and regression.
problem Understanding and improving robustness of neural network predictions.
method Computes reachable sets of neural networks using over- and under-approximations.
result Approach outperforms adversarial attacks and state-of-the-art classifier verification methods.
Deep learning method proves convergence for solving HJI equations.
problem Solving Hamilton-Jacobi-Isaacs equations for reachability analysis.
method Uniform convergence guarantee for DeepReach algorithm.
result DeepReach algorithm converges to classical solution of HJI equation.
Study of 2D Lorentzian anti-de Sitter plane using geometric control theory.
problem Understanding extremal trajectories and reachable set on anti-de Sitter plane.
method Geometric control theory and differential geometry.
result Construction of optimal synthesis and description of Lorentzian distance.
Improves data-driven reachability estimation for complex systems.
problem Estimating reachable states in complex dynamical systems with unknown parameters.
method Uses Christoffel functions and conformal prediction to improve sample efficiency and robustness.
result Guaranteed convergence to the true reach set with improved sample efficiency and robustness.
Quantum neural networks need both data-dependent and trainable unitaries for effective geometric deformation.
problem Quantum neural networks lack the geometric flexibility of classical networks due to limitations in state reachability.
method Viewing quantum states as embedded manifolds, we analyze infinitesimal unitary actions and introduce the CLA maps and aCLS criterion.
result Geometric flexibility in quantum neural networks requires a joint dependence on data and trainable weights.
In this paper, we consider two cases of rolling of one smooth connected complete Riemannian manifold ( M , g ) (M,g) ( M , g ) onto another one $(\hM,\hg)$ of equal dimension n ≥ 2 n\geq 2 n ≥ 2 . The rolling problem ( N S ) (NS) ( N S ) corresponds to the situation where there is no relative spin (or twist) of one manifold with respect to the other one. As for…
The study proposes an audit to assess user control over recommendations in collaborative filtering systems.
problem The gap between maximizing accuracy and ensuring user control over information availability in recommender systems.
method The approach involves a computationally efficient audit for top- N N N linear recommender models, focusing on reachability and user agency. result The study demonstrates that model complexity affects the effort required for users to exert control over their recommendations.
Compact reachability embeddings for hierarchical data
problem Computing geometric representations of hierarchical data
method Using embeddings to represent hierarchies
result Proven compact embeddings for directed trees and graphs of treewidth
Survey of methods to verify deep neural networks.
problem Challenges in verifying deep neural networks for specific properties.
method Borrowing insights from reachability analysis, optimization, and search.
result Comparison of existing algorithms and implementations.
CAST improves spectral clustering for multi-scale data by integrating reachability similarity.
problem Applying spectral clustering to multi-scale data where clusters vary in size and density.
method CAST integrates reachability similarity with distance-based similarity to derive a coefficient matrix, then applies trace Lasso regularization.
result CAST provides excellent performance and robustness across various multi-scale data test cases.
This work explores how feature decorrelation improves self-supervised learning.
problem Complete and dimensional collapse issues in self-supervised learning.
method Study of a concise framework and connection to feature decorrelation.
result Feature decorrelation improves self-supervised learning.
NSC classifies hybrid system states for time-bounded reachability, achieving high accuracy with minimal false negatives.
problem Classifying states in hybrid systems for time-bounded reachability.
method Neural State Classification using Deep Neural Networks.
result Achieved 99.25% to 99.98% accuracy with false-negative rates reduced to 0.0015 to 0 after tuning.
Improved exploration algorithm for unknown MDPs with reduced sample complexity.
problem Exploration of unknown environments without reward function.
method Incremental model-based approach that interleaves state discovery and policy improvement.
result Achieves sample complexity scaling as O ~ ( L 5 S L + ε Γ L + ε A ε − 2 ) \tilde{O}(L^5 S_{L+ε} Γ_{L+ε} A ε^{-2}) O ~ ( L 5 S L + ε Γ L + ε A ε − 2 ) . In finance industry portfolio construction deals with how to divide the investors' wealth across an asset-classes' menu in order to maximize the investors' gain. Main approaches in use at the present are based on variations of the classical Markowitz model. However, recent evolutions of the world market showed limitati…
New distances for causal graphs improve evaluation of learned structures.
problem Difficulty in evaluating graphs learned by causal discovery algorithms.
method Developed a framework for causal distances, including new reachability algorithms.
result Improved distances are faster and more scalable than existing methods.
Study on rolling of 2D and 3D manifolds, identifying orbit dimensions.
problem Understanding rolling dynamics of 2D and 3D manifolds with constraints.
method Modeling rolling as a control affine system on a fibered space Q, analyzing reachable sets.
result Identified possible dimensions of non-open rolling orbits: 2, 5, 6, 7.
We construct normal forms for Lorentzian metrics on Engel distributions under the assumption that abnormal curves are timelike future directed Hamiltonian geodesics. Then we indicate some cases in which the abnormal timelike future directed curve initiating at the origin is geometrically optimal. We also give certain e…
Safe imitation learning with a safety layer for flexible training.
problem Flexible yet safe imitation learning for complex tasks.
method Theory and modular method with a safety layer for continuous policy, adversarial training, and worst-case safety guarantees.
result Robustness advantage of safety layer during training compared to test time.
A new clustering algorithm considers data smoothness for better performance.
problem Clustering multi-scale data with varying cluster densities.
method Divide objects into tiny clusters, cluster centers form smooth graphs.
result Significantly outperforms state-of-the-art clustering algorithms.
In classical reinforcement learning, when exploring an environment, agents accept arbitrary short term loss for long term gain. This is infeasible for safety critical applications, such as robotics, where even a single unsafe action may cause system failure. In this paper, we address the problem of safely exploring fin…
CUDC collects diverse data for offline RL by predicting future states.
problem Challenges in collecting task-agnostic data for offline RL.
method Adaptive temporal distances for curiosity-driven data collection.
result CUDC outperforms existing unsupervised methods in offline RL tasks.
Improves RL from historical data by stitching trajectories.
problem Lack of high-quality data for offline RL.
method Trajectory Stitching (TS) to augment historical data with synthetic actions.
result Improves RL policy performance over baseline.
We consider examples of the H \mathbb H H -type groups with the natural horizontal distribution generated by the commutation relations of the group. In the contrast with the previous studies we furnish the horizontal distribution with the Lorentzian metric, which is nondegenerate metric of index 1 instead of a positive de…
GoTube verifies neural networks over time, scaling to large horizons.
problem Verifying the robustness of time-continuous neural networks.
method Solves Go problems to construct a conservative execution set.
result Substantially outperforms existing tools in size, speed, and scalability.
C-Learning estimates reachability over time to solve multi-goal tasks.
problem Multi-goal reaching challenges in reinforcement learning.
method Cumulative accessibility functions and recurrence relations.
result Optimal cumulative accessibility functions are monotonic in horizon.
DARLING tackles non-stationary RL with guarantees, improving dynamic regret.
problem Non-stationary reinforcement learning in unknown change points.
method Detection Augmented Reinforcement Learning (DARLING) for tabular and linear MDPs.
result DARLING matches minimax lower bounds in tabular and linear MDPs.
New loss function improves neural network robustness.
problem Improving neural network robustness to adversarial attacks.
method Combining energy landscape techniques with a novel loss function.
result The proposed loss function leads to more robust neural networks without significant degradation in classification error.
Understanding efficiency in high dimensional linear models is a longstanding problem of interest. Classical work with smaller dimensional problems dating back to Huber and Bickel has illustrated the benefits of efficient loss functions. When the number of parameters p p p is of the same order as the sample size n n n , $p \…
Study improves understanding of why agentic theorem provers succeed.
problem Understanding which components of agentic theorem provers improve proof success.
method Statistical provability theory and finite-horizon reachability MDP model.
result Bounds provability gap and explains components' effectiveness.
A new curiosity method uses episodic memory to reward novelty, improving learning from sparse rewards.
problem Sparse rewards in real-world reinforcement learning.
method Uses episodic memory to form novelty bonuses based on reachability.
result Improves learning from sparse rewards in various environments.
The paper certifies neural network-based control barrier functions efficiently.
problem Certifying neural network-based barrier functions for safety in autonomous systems.
method Combines NN reachability and hyperplane arrangement enumeration for efficient certification.
result Soundly finds regions where neural networks are certified as barrier functions.
We prove a lower bound for feature dimension in linear MDPs and propose a novel dynamics aggregation framework.
problem The limitation of feature dimension in linear MDPs and the need for efficient hierarchical reinforcement learning.
method We propose a novel dynamics aggregation framework based on structural dynamics and design a provably efficient hierarchical reinforcement learning algorithm.
result Our algorithm achieves a regret of i l d e O ( d ψ 3 / 2 H 3 / 2 N T ) ilde{O} ( d_ψ^{3/2} H^{3/2}\sqrt{ N T} ) i l d e O ( d ψ 3/2 H 3/2 N T ) and meets the condition d ψ 3 N ≪ d 3 d_ψ^3 N \ll d^{3} d ψ 3 N ≪ d 3 in most real-world environments. Safe learning for optimal control with known and unknown dynamics.
problem Safe learning of control strategies for systems with unknown dynamics.
method Reachability analysis and Gaussian Process regression for updating disturbances.
result Algorithm learns optimal control policies without violating safety constraints.
ViTaX provides formal guarantees for targeted explanations in safety-critical systems.
problem Need trustworthy explanations for safety-critical deep neural networks.
method Formal reachability analysis for targeted, semifactual explanations.
result First method to provide formally guaranteed explanations of model resilience.