Complex, temporally evolving phenomena, from climate to brain activity, are governed by dynamical systems (DS). DS reconstruction (DSR) seeks to infer generative surrogate models of these from observed data, reproducing their long-term behavior. Existing DSR approaches require purpose-training for any new system observed, lacking the zero-shot and in-context inference capabilities known from LLMs. Here we introduce DynaMix, a novel multivariate ALRNN-based mixture-of-experts architecture pre-trained for DSR, the first DSR model able to generalize zero-shot to out-of-domain DS. Just from a provided context signal, without any re-training, DynaMix faithfully forecasts the long-term evolution of novel DS where existing time series (TS) foundation models, like Chronos, fail -- at a fraction of the number of parameters (0.1%) and orders of magnitude faster inference times. DynaMix outperforms TS foundation models in terms of long-term statistics, and often also short-term forecasts, even on real-world time series, like traffic or weather data, typically used for training and evaluating TS models, but not at all part of DynaMix' training corpus. We illustrate some of the failure modes of TS models for DSR problems, and conclude that models built on DS principles may bear a huge potential also for advancing the TS prediction field.

Imagine hearing a dog bark and turning toward the sound only to see a parked car, while the real, silent dog sits elsewhere. Such sensory conflicts test perception, yet humans reliably resolve them by prioritizing sound over misleading visuals. Despite advances in multimodal AI integrating vision and audio, little is known about how these systems handle cross-modal conflicts or whether they favor one modality. In this study, we systematically examine modality bias and conflict resolution in AI sound localization. We assess leading multimodal models and benchmark them against human performance in psychophysics experiments across six audiovisual conditions, including congruent, conflicting, and absent cues. Humans consistently outperform AI, demonstrating superior resilience to conflicting or missing visuals by relying on auditory information. In contrast, AI models often default to visual input, degrading performance to near chance levels. To address this, we propose a neuroscience-inspired model, EchoPin, which uses a stereo audio-image dataset generated via 3D simulations. Even with limited training data, EchoPin surpasses existing benchmarks. Notably, it also mirrors human-like horizontal localization bias favoring left-right precision-likely due to the stereo audio structure reflecting human ear placement. These findings underscore how sensory input quality and system architecture shape multimodal representation accuracy.

This paper introduces an automated framework WSW2.0 for analyzing vocal interactions in preschool classrooms, enhancing both accuracy and scalability through the integration of wav2vec2-based speaker classification and Whisper (large-v2 and large-v3) speech transcription. A total of 235 minutes of audio recordings (160 minutes from 12 children and 75 minutes from 5 teachers), were used to compare system outputs to expert human annotations. WSW2.0 achieves a weighted F1 score of .845, accuracy of .846, and an error-corrected kappa of .672 for speaker classification (child vs. teacher). Transcription quality is moderate to high with word error rates of .119 for teachers and .238 for children. WSW2.0 exhibits relatively high absolute agreement intraclass correlations (ICC) with expert transcriptions for a range of classroom language features. These include teacher and child mean utterance length, lexical diversity, question asking, and responses to questions and other utterances, which show absolute agreement intraclass correlations between .64 and .98. To establish scalability, we apply the framework to an extensive dataset spanning two years and over 1,592 hours of classroom audio recordings, demonstrating the framework's robustness for broad real-world applications. These findings highlight the potential of deep learning and natural language processing techniques to revolutionize educational research by providing accurate measures of key features of preschool classroom speech, ultimately guiding more effective intervention strategies and supporting early childhood language development.

Measuring scientific paper innovation is both important and challenging. Existing content-based methods often overlook the full-paper context, fail to capture the full scope of innovation, and lack generalization. We propose HSPIM, a hierarchical and training-free framework based on large language models (LLMs). It introduces a Paper-to-Sections-to-QAs decomposition to assess innovation. We segment the text by section titles and use zero-shot LLM prompting to implement section classification, question-answering (QA) augmentation, and weighted innovation scoring. The generated QA pair focuses on section-level innovation and serves as additional context to improve the LLM scoring. For each chunk, the LLM outputs a novelty score and a confidence score. We use confidence scores as weights to aggregate novelty scores into a paper-level innovation score. To further improve performance, we propose a two-layer question structure consisting of common and section-specific questions, and apply a genetic algorithm to optimize the question-prompt combinations. Furthermore, under the fine-grained structure of innovation, we extend HSPIM to an HSPIM$^+$ that generates novelty, contribution, and feasibility scores with respective confidence scores. Comprehensive experiments on scientific conference paper datasets show that HSPIM outperforms baseline methods in effectiveness, generalization, and interpretability. Demo code is available at https://github.com/Jasaxion/HSPIM.

We introduce MLRC-Bench, a benchmark designed to quantify how effectively language agents can tackle challenging Machine Learning (ML) Research Competitions, with a focus on open research problems that demand novel methodologies. Unlike prior work, e.g., AI Scientist, which evaluates the end-to-end agentic pipeline by using LLM-as-a-judge, MLRC-Bench measures the key steps of proposing and implementing novel research methods and evaluates them with rigorous protocol and objective metrics. Our curated suite of 7 competition tasks reveals significant challenges for LLM agents. Even the best-performing tested agent (gemini-exp-1206 under MLAB) closes only 9.3% of the gap between baseline and top human participant scores. Furthermore, our analysis reveals a misalignment between the LLM-judged innovation and actual performance on cutting-edge ML research problems. MLRC-Bench is a dynamic benchmark, designed to grow with new ML competitions and encourage rigorous, objective evaluations of AI research capabilities. Our leaderboard and code are available at: https://huggingface.co/spaces/launch/MLRC_Bench

Vision Transformers (ViTs) have achieved state-of-the-art performance across various computer vision tasks, but their high computational cost remains a challenge. Token pruning has been proposed to reduce this cost by selectively removing less important tokens. While effective in vision tasks by discarding non-object regions, applying this technique to audio tasks presents unique challenges, as distinguishing relevant from irrelevant regions in time-frequency representations is less straightforward. In this study, for the first time, we applied token pruning to ViT -based audio classification models using Mel-spectrograms and analyzed the trade-offs between model performance and computational cost: TopK token pruning can reduce MAC operations of AudioMAE and AST by 30-40%, with less than a 1% drop in accuracy. Our analysis reveals that while high-intensity or high-variation tokens contribute significantly to model accuracy, low-intensity or low-variation tokens also remain important when token pruning is applied; pruning solely based on the intensity or variation of signals in a patch leads to a noticeable drop in accuracy. We support our claim by measuring high correlation between attention scores and these statistical features and by showing retained tokens consistently receive distinct attention compared to pruned ones. We also show that AudioMAE retains more low-intensity tokens than AST. This can be explained by AudioMAE's self-supervised reconstruction objective, which encourages attention to all patches, whereas AST's supervised training focuses on label-relevant tokens.

The team formation problem assumes a set of experts and a task, where each expert has a set of skills and the task requires some skills. The objective is to find a set of experts that maximizes coverage of the required skills while simultaneously minimizing the costs associated with the experts. Different definitions of cost have traditionally led to distinct problem formulations and algorithmic solutions. We introduce the unified TeamFormation formulation that captures all cost definitions for team formation problems that balance task coverage and expert cost. Specifically, we formulate three TeamFormation variants with different cost functions using quadratic unconstrained binary optimization (QUBO), and we evaluate two distinct general-purpose solution methods. We show that solutions based on the QUBO formulations of TeamFormation problems are at least as good as those produced by established baselines. Furthermore, we show that QUBO-based solutions leveraging graph neural networks can effectively learn representations of experts and skills to enable transfer learning, allowing node embeddings from one problem instance to be efficiently applied to another.

Nowadays, there has been a growing trend in the field of high-energy physics (HEP), in both its experimental and phenomenological studies, to incorporate machine learning (ML) and its specialized branch, deep learning (DL). This review paper provides a thorough illustration of these applications using different ML and DL approaches. The first part of the paper examines the basics of various particle physics types and establishes guidelines for assessing particle physics alongside the available learning models. Next, a detailed classification is provided for representing Jets that are reconstructed in high-energy collisions, mainly in proton-proton collisions at well-defined beam energies. This section covers various datasets, preprocessing techniques, and feature extraction and selection methods. The presented techniques can be applied to future hadron-hadron colliders (HHC), such as the high-luminosity LHC (HL-LHC) and the future circular collider - hadron-hadron (FCChh). The authors then explore several AI techniques analyses designed specifically for both image and point-cloud (PC) data in HEP. Additionally, a closer look is taken at the classification associated with Jet tagging in hadron collisions. In this review, various state-of-the-art (SOTA) techniques in ML and DL are examined, with a focus on their implications for HEP demands. More precisely, this discussion addresses various applications in extensive detail, such as Jet tagging, Jet tracking, particle classification, and more. The review concludes with an analysis of the current state of HEP using DL methodologies. It highlights the challenges and potential areas for future research, which are illustrated for each application.

Large Language Models (LLMs) are used for Register-Transfer Level (RTL) code generation, but they face two main challenges: functional correctness and Power, Performance, and Area (PPA) optimization. Iterative, feedback-based methods partially address these, but they are limited to local search, hindering the discovery of a global optimum. This paper introduces REvolution, a framework that combines Evolutionary Computation (EC) with LLMs for automatic RTL generation and optimization. REvolution evolves a population of candidates in parallel, each defined by a design strategy, RTL implementation, and evaluation feedback. The framework includes a dual-population algorithm that divides candidates into Fail and Success groups for bug fixing and PPA optimization, respectively. An adaptive mechanism further improves search efficiency by dynamically adjusting the selection probability of each prompt strategy according to its success rate. Experiments on the VerilogEval and RTLLM benchmarks show that REvolution increased the initial pass rate of various LLMs by up to 24.0 percentage points. The DeepSeek-V3 model achieved a final pass rate of 95.5\%, comparable to state-of-the-art results, without the need for separate training or domain-specific tools. Additionally, the generated RTL designs showed significant PPA improvements over reference designs. This work introduces a new RTL design approach by combining LLMs' generative capabilities with EC's broad search power, overcoming the local-search limitations of previous methods.

Test-time scaling methods have seen a rapid increase in popularity for its computational efficiency and parameter-independent training to improve reasoning performance on Large Language Models. One such method is called budget forcing, a decoding intervention strategy which allocates extra compute budget for thinking and elicits the inherent self-correcting behavior of the model. However, this relies on supervised fine-tuning (SFT) on long-context reasoning traces which causes performance degradation on smaller models due to verbose responses. For this reason, we offer a framework integrating reinforcement learning (RL) to improve token efficiency and boost the performance of a 1.5B model for mathematical reasoning. We demonstrate this using only 1.5K training samples and found that our SFT+RL model performed better on the GSM8K dataset with varying compute budgets. Our main findings showed an overall higher accuracy while significantly reducing its token usage by over 40% compared to the SFT model, revealing how RL can recover the losses due to long-context training and altogether improving performance in mathematical reasoning.