Machine learning models excel with abundant annotated data, but annotation is often costly and time-intensive. Active learning (AL) aims to improve the performance-to-annotation ratio by using query methods (QMs) to iteratively select the most informative samples. While AL research focuses mainly on QM development, the evaluation of this iterative process lacks appropriate performance metrics. This work reviews eight years of AL evaluation literature and formally introduces the speed-up factor, a quantitative multi-iteration QM performance metric that indicates the […]

Ver mais

arXiv:2604.06265v1 Announce Type: new Abstract: Quantum-inspired tensor networks algorithms have shown to be effective and efficient models for machine learning tasks, including anomaly detection. Here, we propose a highly parallelizable quantum-inspired approach which we call SMT-AD from Superposition of Multiresolution Tensors for Anomaly Detection. It is based upon the superposition of bond-dimension-1 matrix product operators to transform the input data with Fourier-assisted feature embedding, where the number of learnable parameters grows linearly with feature size, embedding resolutions, and […]

Ver mais

Cosine similarity is prevalent in contrastive learning, yet it makes an implicit assumption: embedding magnitude is noise. Prior work occasionally found dot product and cosine similarity comparable, but left unanswered WHAT information magnitude carries, WHEN it helps, and HOW to leverage it. We conduct a systematic study through a $2 times 2$ ablation that independently controls input-side and output-side normalization across text and vision models. Our findings reveal three key insights. First, in text retrieval, output (document) magnitude […]

Ver mais

Was Daft Punk Having a Laugh When They Chose the Tempo of Harder, Better, Faster, Stronger? Depending on how you measure it, the tempo of Harder, Better, Faster, Stronger appears to be 123.45 beats per minute. This is one of those things that’s so cool I’m just going to accept it as true. (I only today learned from the Hacker News comments that Veridis Quo is “Very Disco”, and if you flip the order of those words you […]

Ver mais

We introduce Aumann-SHAP, an interaction-aware framework that decomposes counterfactual transitions by restricting the model to a local hypercube connecting baseline and counterfactual features. Each hyper-cube is decomposed into a grid in order to construct an induced micro-player cooperative game in which elementary grid-step moves become players. Shapley and LES values on this TU-micro-game yield: (i) within-pot contribution of each feature to the interaction with other features (interaction explainability), and (ii) the contribution of each instance and each feature […]

Ver mais

Recent progress in reinforcement learning with verifiable rewards (RLVR) shows that small, specialized language models (SLMs) can exhibit structured reasoning without relying on large monolithic LLMs. We introduce soft hidden-state collaboration, where multiple heterogeneous frozen SLM experts are integrated through their internal representations via a trainable attention interface. Experiments on Reasoning Gym and GSM8K show that this latent integration is competitive with strong single-model RLVR baselines. Ablations further reveal a dual mechanism of expert utilization: for simpler arithmetic […]

Ver mais

Human feedback often arrives as preferences rather than calibrated numeric rewards, motivating reinforcement learning from preferential feedback, also referred to as reinforcement learning from human feedback (RLHF). We present a rigorous theoretical study of preference-only learning in episodic kernel MDPs. In each episode, the learner deploys two policies from a common start state and receives a single binary label indicating which trajectory is preferred, modeled by a Bradley–Terry–Luce link on the difference of cumulative (unobserved) rewards. Under kernel-based […]

Ver mais

arXiv:2603.00555v1 Announce Type: new Abstract: Skill-based programming of robots provides a flexible approach for automation. Existing solutions neglect the optimization of motion sequences, leading to inefficiencies in execution. This work introduces a planning method that enhances skill-based robot programming by integrating motion sequence optimization. This optimization leads to a new MoveContinuousSkill. The software for executing the MoveContinuousSkill is implemented on a Programmable Logic Controller and applied across multiple robotic systems. Experimental results demonstrate a significant improvement in execution […]

Ver mais