arXiv research

pLSTM tackles long-range language modeling and computer vision tasks with parallelizable linear source transition mark networks.

In the following paper we investigate the question: when is a transitive topological groupoid continuously isomorphic to a Lie groupoid? We present many results on the matter which may be considered generalizations of the Hilbert's fifth problem to this context. Most notably we present a "solution" to the problem for p…

In the following paper we investigate the question: when is a transitive topological groupoid continuously isomorphic to a Lie groupoid? We present many results on the matter which may be considered generalizations of the Hilbert's fifth problem to this context. Most notably we present a "solution" to the problem for p…

Paper analyzes latent space geometry in generative models using Fisher information.

The paper analyzes phase transitions in transfer learning for perceptrons.

Let G be a Lie groupoid over M such that the target-source map from G to M x M is proper. We show that, if O is an orbit of finite type (i.e. which admits a proper function with finitely many critical points), then the restriction G|U of G to some neighborhood U of O in M is isomorphic to a similar restriction of the a…

Neural networks with DAGs show linearity as width increases.

A new method for reinforcement learning that adapts to different domains using auxiliary classifiers.

A new framework for robust policy learning in MDPs with linear mixture dynamics.

SCPO learns robust policies without modeling disturbance, improving real-world task performance.

The paper investigates non-linear and heavy-tailed predictability in transition-energy financial markets.

New algorithm reduces reinforcement learning regret for linear MDPs with unknown transitions.

The notion of $n$-transitivity can be carried over from groups of diffeomorphisms on a manifold $M$ to groups of bisections of a Lie groupoid over $M$. The main theorem states that the $n$-transitivity is fulfilled for all $n\in\mathbb N$ by an arbitrary group of $C^r$-bisections of a Lie groupoid $Γ$ of class $C^r$, w…

Study of transitivity in partially hyperbolic maps with expanding linear part.

RFMs transition from linear to nonlinear under specific input-label correlation.

For many years, a combination of principal component analysis (PCA) and independent component analysis (ICA) has been used for blind source separation (BSS). However, it remains unclear why these linear methods work well with real-world data that involve nonlinear source mixtures. This work theoretically validates that…

A new method for ILO with transition model disparity using an intermediary policy.

The paper tackles joint learning of linear systems, improving accuracy with pooled data.

Neural models learn continuous-time Markov chain transition rates from data.

Analyzes bias-variance in overparameterized linear models using random features.

In most sampling algorithms, including Hamiltonian Monte Carlo, transition rates between states correspond to the probability of making a transition in a single time step, and are constrained to be less than or equal to 1. We derive a Hamiltonian Monte Carlo algorithm using a continuous time Markov jump process, and ar…

UCB-TQL learns from multiple tasks with shared dynamics and adapts to task-specific variations.

We describe a generalization of the Hierarchical Dirichlet Process Hidden Markov Model (HDP-HMM) which is able to encode prior information that state transitions are more likely between "nearby" states. This is accomplished by defining a similarity function on the state space and scaling transition probabilities by pai…

Wide neural networks become linear, but adding bottlenecks makes them bilinear or multilinear.

StrADiff separates sources from mixtures without labels, using structured priors.

The paper analyzes how combining samples from two tasks can improve performance, especially in high dimensions.

Within the last two decades, Foreign Direct Investment (FDI) has been observed as one of the prime instruments in the process of restructuring the European economies in transition. Many scholars argue that FDI is expected to be a source of valuable technology transfer thus might certainly have positive effects on host …

We address the problem of necessary conditions and topological obstructions for the existence of robustly transitive maps on surfaces. Concretely, we show that partial hyperbolicity is a necessary condition in order to have $C^1$ robustly transitive endomorphisms with critical points on surfaces, and the only surfaces …

Study shows optimal RL with transition look-ahead is NP-hard for $\ell \geq 2$.

SVM and linear regression models coincide in high dimensions.

New algorithm reduces dynamic regret for MDPs with unknown transition and adversarial rewards.

We give a polynomial-time algorithm for learning latent-state linear dynamical systems without system identification, and without assumptions on the spectral radius of the system's transition matrix. The algorithm extends the recently introduced technique of spectral filtering, previously applied only to systems with a…

The waggle dance that honeybees perform is an astonishing way of communicating the location of food source. After over 60 years of its discovery, researchers still use manual labeling by watching hours of dance videos to detect different transitions between dance components thus extracting information regarding the dis…

Wide neural networks become linear, with constant tangent kernel, due to Hessian scaling.

The energy transition is well underway in most European countries. It has a growing impact on electric power systems as it dramatically modifies the way electricity is produced. To ensure a safe and smooth transition towards a pan-European electricity production dominated by renewable sources, it is of paramount import…

Paper tackles robust decision-making from multiple sites with shared structure.

We present a variational formulation of electrodynamics using de Rham even and odd differential forms. Our formulation relies on a variational principle more complete than the Hamilton principle and thus leads to field equations with external sources and permits the derivation of the constitutive relations. We interpre…

Study uses Wasserstein distance to identify causal orders and unmix sources.

Deep reinforcement learning has recently shown many impressive successes. However, one major obstacle towards applying such methods to real-world problems is their lack of data-efficiency. To this end, we propose the Bottleneck Simulator: a model-based reinforcement learning method which combines a learned, factorized …

This research improves dynamical systems understanding by identifying latent states and their nonlinear transitions.

This paper studies the dynamic generator model for spatial-temporal processes such as dynamic textures and action sequences in video data. In this model, each time frame of the video sequence is generated by a generator model, which is a non-linear transformation of a latent state vector, where the non-linear transform…

We present a simple model of firm rating evolution. We consider two sources of defaults: individual dynamics of economic development and Potts-like interactions between firms. We show that such a defined model leads to phase transition, which results in collective defaults. The existence of the collective phase depends…

We consider the problem of learning in Linear Quadratic Control systems whose transition parameters are initially unknown. Recent results in this setting have demonstrated efficient learning algorithms with regret growing with the square root of the number of decision steps. We present new efficient algorithms that ach…

Efficient RL for linear MDPs with unknown transitions.

The paper estimates key metrics for linear models with Markov or hidden Markov sources.

The paradox of the energy transition is that the low marginal costs of new renewable energy sources (RES) drag electricity prices down and discourage investments in flexible productions that are needed to compensate for the lack of dispatchability of the new RES. The energy transition thus discourages the investments t…

Unsupervised Domain Adaptation (DA) is used to automatize the task of labeling data: an unlabeled dataset (target) is annotated using a labeled dataset (source) from a related domain. We cast domain adaptation as the problem of finding stable labels for target examples. A new definition of label stability is proposed, …

A new parallel algorithm speeds up Hawkes process estimation.