arXiv research

Optimally explores dynamical systems with varying properties using context inference.

A method for learning a context latent vector to improve generalization in model-based RL.

CoDA adapts dynamics models to new physical systems by conditioning on context.

New framework handles dynamic contexts in reinforcement learning.

Efficiently selects top-m designs for various contexts using sequential sampling.

Transformers learn rich in-context dependencies efficiently.

Proposes a new model for more accurate demand forecasting considering dynamic contextual information.

LC-SAC tackles non-stationary dynamics in reinforcement learning.

Paper analyzes Transformer learning dynamics, proving benign landscape for in-context learning.

Attention-only transformers learn from context via two stages of inference.

Transformers achieve near-optimal dynamic regret in non-stationary reinforcement learning.

Market-GAN adds context control to financial market data generation.

The paper tackles long-context linear system identification with improved sample complexity bounds.

HiPPO-Prophecy models can learn dynamical systems without fine-tuning.

In this paper, we put the issue of dynamic equivalence of control systems in the context of pullbacks of coframings on infinite jet bundles over the state manifolds. While much attention has been given to differentially flat systems, i.e. systems dynamically equivalent to linear control systems, the advantage of this a…

Improved algorithms for dynamic pricing under different valuation models.

Transformers can learn noisy linear systems with depth and IID data.

A Riemmanian foliated dynamical system of 3-dimension $(\mathrm{RFDS}^{3})$ is a closed Riemannian 3-manifold with additional structures: foliation, dynamical system. In the context of arithmetic topology, it is a geometric/analytic analogue of an arithmetic scheme with a conjectural dynamical system suggested by C. De…

Recent years have witnessed the world-wide emergence of mega-metropolises with incredibly huge populations. Understanding residents mobility patterns, or urban dynamics, thus becomes crucial for building modern smart cities. In this paper, we propose a Neighbor-Regularized and context-aware Non-negative Tensor Factoriz…

VectorNet predicts car behavior using vectorized HD maps and agent dynamics.

Study on multi-head softmax attention dynamics for in-context learning.

Transformers can learn mixture of linear models efficiently.

DCE learns customer embeddings from digital activity and financial context.

Paper tackles overfitting in RL for trade execution.

Transformers can implement reinforcement learning algorithms from data without updates.

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

In the context of an incomplete market with a Brownian filtration and a fixed finite time horizon, this paper proves that for general dynamic convex risk measures, the buyer's and seller's risk indifference prices of a contingent claim are bounded from below and above by the dynamic lower and upper hedging prices, resp…

Intratumor heterogeneity is often manifested by vascular compartments with distinct pharmacokinetics that cannot be resolved directly by in vivo dynamic imaging. We developed tissue-specific compartment modeling (TSCM), an unsupervised computational method of deconvolving dynamic imaging series from heterogeneous tumor…

Transformers learn unseen tasks via prompts without fine-tuning.

Novel framework for systemic risk analysis in financial markets.

Designing optimal controllers continues to be challenging as systems are becoming complex and are inherently nonlinear. The principal advantage of reinforcement learning (RL) is its ability to learn from the interaction with the environment and provide optimal control strategy. In this paper, RL is explored in the cont…

Study $L_2$ regularization in deep networks, uncovering performance relations and proposing a training schedule.

Transformers struggle to learn Markovian dynamics, showing NP-hard optimization challenges.

Recent advancements in recurrent neural network (RNN) research have demonstrated the superiority of utilizing multiscale structures in learning temporal representations of time series. Currently, most of multiscale RNNs use fixed scales, which do not comply with the nature of dynamical temporal patterns among sequences…

We consider a context-based dynamic pricing problem of online products, which have low sales. Sales data from Alibaba, a major global online retailer, illustrate the prevalence of low-sale products. For these products, existing single-product dynamic pricing algorithms do not work well due to insufficient data samples.…

The Dynamic Chain Event Graph (DCEG) is able to depict many classes of discrete random processes exhibiting asymmetries in their developments and context-specific conditional probabilities structures. However, paradoxically, this very generality has so far frustrated its wide application. So in this paper we develop an…

EFA extends self-attention to handle mixed data types and dynamic relevance.

New attack reveals memorization patterns in pre-trained LLMs.

Proposes a dynamic matching algorithm for two-sided online markets.

Transformers learn algorithms for in-context learning with bounds and stability analysis.

This short expository note gives an elementary introduction to the study of dynamics on certain moduli spaces, and in particular the recent breakthrough result of Eskin, Mirzakhani, and Mohammadi. We also discuss the context and applications of this result, and connections to other areas of mathematics such as algebrai…

We introduce dynamic asymptotic dimension, a notion of dimension for actions of discrete groups on locally compact spaces, and more generally for locally compact étale groupoids. We study our notion for minimal actions of the integer group, its relation with conditions used by Bartels, Lück, and Reich in the context of…

Uncertainty modeling for dynamical systems

Particles representing tokens cluster in Transformers, influenced by initial tokens and matrix spectrum.

asp2vec learns dynamic node aspect distributions for better network embedding.

The study identifies and analyzes different market regimes in equity markets using advanced signal processing techniques.

This paper approaches the definition and properties of dynamic convex risk measures through the notion of a family of concave valuation operators satisfying certain simple and credible axioms. Exploring these in the simplest context of a finite time set and finite sample space, we find natural risk-transfer and time-co…

Direct approach for handling contextual bandits with latent state dynamics.