We consider a learning augmented framework for NP-hard permutation problems. The algorithm has access to predictions telling, given a pair $u,v$ of elements, whether $u$ is before $v$ or not in an...

In this work, we investigate the merits of explicitly optimizing for inference time algorithmic performance during model training. We show how optimizing for inference time performance can improve...

Large language models (LLMs) deployed as agents solve user-specified tasks over multiple steps while keeping the required manual engagement to a minimum. Crucially, such LLMs need to ground their...

As large language models increasingly drive real-world applications, aligning them with human values becomes paramount. Reinforcement Learning from Human Feedback (RLHF) has emerged as a key...

Circular and non-flat data distribution are prevalent across diverse domains of data science, yet their specific geometric structures often remain underutilized in machine learning frameworks. A...

This paper introduces von Mises-Fisher exploration (vMF-exp), a scalable method for exploring large action sets in reinforcement learning problems where hyperspherical embedding vectors represent...

Increasing diversity in language models is a challenging yet essential objective. A common approach is to raise the decoding temperature. In this work, we investigate this approach through a...

Traditional online file management systems, such as Google Drive and OneDrive, provide convenient platforms for collaborative editing and sharing, but they come with inherent limitations. These systems often offer limited storage space and operate on centralized...

Riemannian methods have established themselves as state-of-the-art approaches in Brain-Computer Interfaces (BCI) in terms of performance. However, their adoption by experimenters is often hindered by a lack of interpretability. In this work, we propose a set of tools designed to enhance practitioners' understanding of the decisions made by Riemannian methods. Specifically, we develop techniques to quantify and visualize the influence of the different sensors on classification outcomes. Our approach includes a visualization tool for high-dimensional covariance matrices, a classifieragnostic tool that focuses on the classification process, as well as methods that leverage the data's topology to better characterize the role of each sensor. We demonstrate these tools on a specific dataset and provide Python code to facilitate their use by practitioners, thereby promoting the adoption of Riemannian methods in BCI.

Let $$G=(V\cup T,E)$$ be a graph where $$V\cup T$$ is the set of vertices, with T a subset of distinguished vertices, called terminals, and E the set of edges. Given a weight function $$w: V\rightarrow \mathbb N{\setminus } \{0\}$$ associated with the nonterminal nodes, the multi-terminal vertex separator problem consists in partitioning $$V\cup T$$ into $$k+1$$ subsets $$\{S, V_1,\dots , V_k\}$$ such that there is no edge between two different subsets $$V_i$$ and $$V_j$$ , each $$V_i$$ contains exactly one terminal and the weight of S is minimum. In this paper, we characterize the polytope of the solutions of this problem for two classes of the graph, and we show that the two linear systems are totally dual integral. Then, we study the polytope for the graphs that are decomposable by 1-node cutsets. We show that if G decomposes into $$G_1, \dots , G_k$$ , then the polytope in G can be obtained from those in $${\bar{G}}_1, \dots , {\bar{G}}_k$$ , where $${\bar{G}}_1, \dots , {\bar{G}}_k$$ are graphs related to $$G_1, \dots , G_k$$ , respectively. We also derive a procedure for composing facets and give some algorithmic consequences for solving the problem in G from $${\bar{G}}_1, \dots , {\bar{G}}_k$$ .

We propose a family a metrics over the set of full-rank $$n\times p$$ real matrices, and apply them to the landing framework for optimization under orthogonality...

The goal of this paper is to show how different machine learning tools on the Riemannian manifold $$\mathcal {P}_d$$ of Symmetric Positive Definite (SPD) matrices...

Object-centric Predictive Monitoring has recently gained attention due to advances in machine learning and rise of Object-Centric Event Logs (OCELs), which comprehensively capture object interactions. This paper presents a modular framework supporting customizable pipelines for predictive analysis across diverse event logs. The framework comprises three core components: Preprocessing (preserving object relationships via graph structures), Graph Embedding Model, and Prediction Model. We experimentally evaluated various combinations of embeddings and predictors on three public OCELs. Results show that no single configuration consistently dominates. However, GAT and Graph Transformer models perform best for predicting remaining time and the number of events. Performance improves with larger embedding and subgraphs, particularly for neuralbased models. Finally, GAT delivered the most stable and highperforming results across all event logs in generalization tests.