Study on parameter dynamics in exponential families under closed-loop learning.
problem Understanding and preventing convergence to biased absorbing states in model parameter estimation.
method Derived equations of motion for parameter dynamics in exponential families. Showed convergence to biased states under maximum likelihood estimation and proposed solutions.
result Closed-loop learning can lead to biased parameter estimates, but can be mitigated by using maximum a posteriori estimation or regularisation.
Framework learns robust control policies from expert demonstrations.
problem Adversarial robustness and closed-loop generalization in feedback control policies.
method Lipschitz-constrained loss minimization for certified robustness and generalization.
result Finite sample bound on policy learning error and robust closed-loop stability.
Formula adjusts steady-state models for control confounding.
problem Learning steady-state models from operational data can be flawed due to control confounding.
method Derives a formula to adjust for control confounding using structural dynamical causal models.
result Estimates a causal steady-state model from closed-loop operational data.
For large-scale industrial processes under closed-loop control, process dynamics directly resulting from control action are typical characteristics and may show different behaviors between real faults and normal changes of operating conditions. However, conventional distributed monitoring approaches do not consider the…
This paper addresses the problem of learning the optimal control policy for a nonlinear stochastic dynamical system with continuous state space, continuous action space and unknown dynamics. This class of problems are typically addressed in stochastic adaptive control and reinforcement learning literature using model-b…
Fault detection problem for closed loop uncertain dynamical systems, is investigated in this paper, using different deep learning based methods. Traditional classifier based method does not perform well, because of the inherent difficulty of detecting system level faults for closed loop dynamical system. Specifically, …
Data-efficient reinforcement learning (RL) in continuous state-action spaces using very high-dimensional observations remains a key challenge in developing fully autonomous systems. We consider a particularly important instance of this challenge, the pixels-to-torques problem, where an RL agent learns a closed-loop con…
Data-efficient learning in continuous state-action spaces using very high-dimensional observations remains a key challenge in developing fully autonomous systems. In this paper, we consider one instance of this challenge, the pixels to torques problem, where an agent must learn a closed-loop control policy from pixel i…
Study bounds the length of shortest periodic geodesics on certain curved spaces.
problem Bounding the length of shortest periodic geodesics on curved spaces.
method Analyzing the space of closed loops and their homotopy.
result The length of a shortest periodic geodesic is bounded by 8π(n−1). Analyzes how learning algorithms affect and are affected by data manipulation.
problem Characterizing the closed-loop behavior of learning algorithms in the presence of decision-dependent data.
method Analyzes repeated risk minimization as perturbed gradient flows of performative risk minimization, considering multiple local minimizers.
result Characterizes the region of attraction for various equilibria and introduces performative alignment.
Geometrically characterizes virtual nonlinear nonholonomic constraints using symplectic methods.
problem Characterizing virtual nonlinear nonholonomic constraints geometrically.
method Geometric characterization using symplectic structures and Chetaev equations.
result A unique control law exists to satisfy virtual constraints, and closed-loop dynamics are projections of uncontrolled dynamics.
Closed loop solitons in a plane, whose curvatures obey the modified Korteweg-de Vries equation, were investigated. It was shown that their tangential vectors are expressed by ratio of Weierstrass sigma functions for genus one case and ratio of Baker's sigma functions for the genus two case. This study is closely relate…
This work discusses a closed-loop control strategy for complex systems utilizing scarce and streaming data. A discrete embedding space is first built using hash functions applied to the sensor measurements from which a Markov process model is derived, approximating the complex system's dynamics. A control strategy is t…
New budget quantifies drift in closed-loop learning, improving reproducibility.
problem Characterizing statistical learning under distributional drift in closed-loop settings.
method Introduces an intrinsic drift budget CT quantifying cumulative information-geometric motion of the data distribution. result Proves a drift-feedback bound of order T−1/2+CT/T for prequential reproducibility, up to controlled second-order remainder terms. CLQT benchmarks LLM portfolio managers by evaluating their decision-making process, not just returns.
problem Most benchmarks rank LLMs by returns, ignoring their decision-making process and potential for look-ahead leakage.
method CLQT reframes evaluation as diagnosis, using a closed-loop, cost-aware, strategy-consistent environment with a five-stage cycle.
result CLQT provides a durable map of agent competencies and limitations, separating outcome from process.
DQN outperforms static policies in a dynamic fee environment for automated market makers.
problem How automated market makers (AMMs) perform under dynamic fees is unknown.
method Constructed a closed-loop simulator with dynamic fees, noise flow, and arbitrage.
result A small DQN policy outperforms static policies in a dynamic fee environment.
Deep neural networks are known to be fragile to small adversarial perturbations. This issue becomes more critical when a neural network is interconnected with a physical system in a closed loop. In this paper, we show how to combine recent works on neural network certification tools (which are mainly used in static set…
Study shows how multiple traders can trade together without excessive price impact.
problem Coordination issues in trading to exploit a common signal.
method Closed-loop Nash competition model for stochastic differential games.
result Excessive trading reduced but not significantly for practical parameters.
The paper tackles performative risk optimization under weak convexity assumptions.
problem Optimizing performative risk in a closed-loop prediction system with weak convexity.
method Relaxing convexity assumptions to maintain optimization feasibility.
result Iterative optimization methods remain applicable even with weakened convexity conditions.
Robot science discovers new materials faster.
problem Discovering advanced materials in complex synthesis landscapes.
method Closed-loop, active learning-driven autonomous system.
result Discovery of a novel epitaxial nanocomposite phase-change memory material.
This paper uses NLDT to find interpretable control rules from complex DRL policies.
problem Complex, non-interpretable policies from black-box AI methods.
method Evolutionary optimization of NLDT for hierarchical control rules.
result Interpretable control rules with similar performance to black-box DRL.
Sequential learning of tasks using gradient descent leads to an unremitting decline in the accuracy of tasks for which training data is no longer available, termed catastrophic forgetting. Generative models have been explored as a means to approximate the distribution of old tasks and bypass storage of real data. Here …
PandaAI: A practical agent for neuro-symbolic data analysis and decision-making in finance
problem Sequential decision-making in finance
method Leveraging LLMs for market regime modeling and constrained alpha generation
result PandaAI achieves higher Rank IC and lower maximum drawdown
We propose a reinforcement learning (RL) based closed loop power control algorithm for the downlink of the voice over LTE (VoLTE) radio bearer for an indoor environment served by small cells. The main contributions of our paper are to 1) use RL to solve performance tuning problems in an indoor cellular network for voic…
A homothety surface can be assembled from polygons by identifying their edges in pairs via homotheties, which are compositions of translation and scaling. We consider linear trajectories on a 1-parameter family of genus-2 homothety surfaces. The closure of a trajectory on each of these surfaces always has Hausdorff dim…
Study finds loops with specific curvature exist using Hardy's inequality.
problem Existence of closed planar loops with prescribed curvature.
method Variational approach, Hardy's inequality and associated functional space.
result Existence of loops with specific curvature proven.
Paper tackles stochastic control with mean and higher-order moments, finding Nash equilibria.
problem Time-inconsistent stochastic control problems with mean and higher-order moments.
method Developed closed-loop and open-loop Nash equilibrium controls using PDEs and maximum principles.
result Identical closed-loop and open-loop Nash equilibria controls, independent of state value and random path.
A geometric approach to differential game theory is illustrated. The parallel pursuit is considered as a two-player zero-sum differential game. The optimal strategies of each player is designed based on Riemann-Finsler geometry. Our approach incorporates a closed loop optimal control and the presentation is familiar wi…
AdaptOn achieves logarithmic regret in adaptive control of unknown partially observable linear systems.
problem Adaptive control in partially observable linear dynamical systems.
method AdaptOn algorithm that estimates system dynamics through online learning and gradient descent.
result AdaptOn achieves a logarithmic regret bound of polylog(T) after T steps.
Bubblewrap predicts neural dynamics online, scaling to thousands of neurons.
problem Direct testing of neural population hypotheses requires online inference of neural state.
method Soft tiling of neural manifold with fast, stable dimensionality reduction.
result Bubblewrap model outperforms existing methods in noisy conditions.
Paper predicts recycling bin full events to reduce RVM downtime.
problem Predicting bin full events to increase RVM uptime.
method Hybrid approach combining machine learning and statistical approximation.
result Forecasting leads to less downtime and costs compared to emptying strategies.
We present an alternative local definition of the writhe of a self-avoiding closed loop which differs from the traditional non-local definition by an integer. When studying dynamics this difference is immaterial. We employ a formula due to Aldinger, Klapper and Tabor for the change in writhe and propose a set of local,…
ApolloRL offers a platform for RL research in autonomous driving.
problem Improving reinforcement learning for autonomous driving.
method Open platform with training, simulation, and evaluation components.
result Baseline agents perform well in the ApolloRL environment.
Filling length measures the length of the contracting closed loops in a null-homotopy. The filling length function of Gromov for a finitely presented group measures the filling length as a function of length of edge-loops in the Cayley 2-complex. We give a bound on the filling length function in terms of the log of an …
Deep RL improves blood glucose control for T1D patients.
problem Managing blood glucose levels for people with type 1 diabetes.
method Developed deep reinforcement learning techniques for automated blood glucose control.
result Deep RL approach outperforms baseline control algorithms, reducing glycemic risk and hypoglycemia.
In this paper, we study the problem of learning vision-based dynamic manipulation skills using a scalable reinforcement learning approach. We study this problem in the context of grasping, a longstanding challenge in robotic manipulation. In contrast to static learning behaviors that choose a grasp point and then execu…
Combines Gaussian processes and polynomial chaos for stochastic control.
problem Uncertainties in dynamic models lead to performance issues in predictive control.
method Combines Gaussian processes with polynomial chaos expansions to estimate probability distributions of nonlinear functions.
result Demonstrates accurate approximation and closed-loop performance in stochastic nonlinear model predictive control.
We prove a theorem about elliptic operators with symmetric potential functions, defined on a function space over a closed loop. The result is similar to a known result for a function space on an interval with Dirichlet boundary conditions. These theorems provide accurate numerical methods for finding the spectra of tho…
We propose a sliding surface for systems on the Lie group SO(3)×R3 . The sliding surface is shown to be a Lie subgroup. The reduced-order dynamics along the sliding subgroup have an almost globally asymptotically stable equilibrium. The sliding surface is used to design a sliding-mode controller for t…
We utilize Wi-Fi communications from smartphones to predict their mobility mode, i.e. walking, biking and driving. Wi-Fi sensors were deployed at four strategic locations in a closed loop on streets in downtown Toronto. Deep neural network (Multilayer Perceptron) along with three decision tree based classifiers (Decisi…
Investigates time-inconsistent portfolio selection under MMV preferences.
problem Time-inconsistent optimal strategies for MMV preferences.
method Nash equilibrium controls for MMV and MV preferences, solving FBSDE and HJB equations.
result MMV optimal strategies lead to higher investment amounts than MV strategies, narrowing over time.
Iterative method learns unknown constraints for MPC control.
problem Learning to satisfy unknown polyhedral state constraints in iterative MPC.
method Collects and improves estimates of unknown constraints using collected data, designs an MPC controller to satisfy the estimated constraints.
result Robust and probabilistic guarantees of constraint satisfaction as a function of task iterations.
Paper develops an AI-driven framework for systematic investing.
problem Manual prompts limit model adaptability and data snooping biases.
method Closed-loop system with self-evolving AI, out-of-sample validation, and economic rationale.
result Long-short portfolios on factor signals outperform with Sharpe ratio 3.11 and return 59.53%.
We show LLMs can be locally linear, enabling better control of activations.
problem Suboptimal control of LLM activations during generation.
method Model LLM inference as a linear dynamical system, compute feedback controllers using Jacobians, and adapt classical control theory.
result Robust, fine-grained control of LLM activations across models and tasks.
Formula for integrating random variables on hyperbolic surfaces.
problem Integrating random variables on the moduli space of hyperbolic surfaces.
method Integration formula for lengths of closed geodesics.
result Integral of geometric random variables can be expressed as an integral over R.
Braidlab is a Matlab package for analyzing data using braids. It was designed to be fast, so it can be used on relatively large problems. It uses the object-oriented features of Matlab to provide a class for braids on punctured disks and a class for equivalence classes of simple closed loops. The growth of loops under …
In this work, we propose a new learning approach for autonomous navigation and landing of an Unmanned-Aerial-Vehicle (UAV). We develop a multimodal fusion of deep neural architectures for visual-inertial odometry. We train the model in an end-to-end fashion to estimate the current vehicle pose from streams of visual an…
This paper reviews recent advances in the field of optimization under uncertainty via a modern data lens, highlights key research challenges and promise of data-driven optimization that organically integrates machine learning and mathematical programming for decision-making under uncertainty, and identifies potential r…