Planning and performing interactive tasks, such as conducting experiments to determine the melting point of an unknown substance, is straightforward for humans but poses significant challenges for autonomous agents. We introduce ReasonPlanner, a novel generalist agent designed for reflective thinking, planning, and interactive reasoning. This agent leverages LLMs to plan hypothetical trajectories by building a World Model based on a Temporal Knowledge Graph. The agent interacts with the environment using a natural language actor-critic module, where the actor translates the imagined trajectory into a sequence of actionable steps, and the critic determines if replanning is necessary. ReasonPlanner significantly outperforms previous state-of-the-art prompting-based methods on the ScienceWorld benchmark by more than 1.8 times, while being more sample-efficient and interpretable. It relies solely on frozen weights thus requiring no gradient updates. ReasonPlanner can be deployed and utilized without specialized knowledge of Machine Learning, making it accessible to a wide range of users.

We present the MahaSUM dataset, a large-scale collection of diverse news articles in Marathi, designed to facilitate the training and evaluation of models for abstractive summarization tas ks in Indic languages. The dataset, containing 25k samples, was create d by scraping articles from a wide range of online news sources and manuall y verifying the abstract summaries. Additionally, we train an IndicBAR T model, a variant of the BART model tailored for Indic languages, usin g the Maha-SUM dataset. We evaluate the performance of our trained mode ls on the task of abstractive summarization and demonstrate their eff ectiveness in producing high-quality summaries in Marathi. Our work cont ributes to the advancement of natural language processing research in Indic languages and provides a valuable resource for future research in this area using state-of-the-art models.

This paper presents the foundational framework and initial findings of the Balanced Arabic Readability Evaluation Corpus (BAREC) project, designed to address the need for comprehensive Arabic language resources aligned with diverse readability levels. Inspired by the Taha/Arabi21 readability reference, BAREC aims to provide a standardized reference for assessing sentence-level Arabic text readability across 19 distinct levels, ranging in targets from kindergarten to postgraduate comprehension. Our ultimate goal with BAREC is to create a comprehensive and balanced corpus that represents a wide range of genres, topics, and regional variations through a multifaceted approach combining manual annotation with AI-driven tools. This paper focuses on our meticulous annotation guidelines, demonstrated through the analysis of 10,631 sentences/phrases (113,651 words). The average pairwise inter-annotator agreement, measured by Quadratic Weighted Kappa, is 79.9%, reflecting a high level of substantial agreement. We also report competitive results for benchmarking automatic readability assessment. We will make the BAREC corpus and guidelines openly accessible to support Arabic language research and education.

Despite the well-established importance of feedback in education, the application of Artificial Intelligence (AI)-generated feedback, particularly from language models like ChatGPT, remains understudied in translation education. This study investigates the engagement of master's students in translation with ChatGPT-generated feedback during their revision process. A mixed-methods approach, combining a translation-and-revision experiment with quantitative and qualitative analyses, was employed to examine the feedback, translations pre-and post-revision, the revision process, and student reflections. The results reveal complex interrelations among cognitive, affective, and behavioural dimensions influencing students' engagement with AI feedback and their subsequent revisions. Specifically, the findings indicate that students invested considerable cognitive effort in the revision process, despite finding the feedback comprehensible. Additionally, they exhibited moderate affective satisfaction with the feedback model. Behaviourally, their actions were largely influenced by cognitive and affective factors, although some inconsistencies were observed. This research provides novel insights into the potential applications of AI-generated feedback in translation teachingand opens avenues for further investigation into the integration of AI tools in language teaching settings.

This paper presents a theoretical framework for an AI-driven data quality monitoring system designed to address the challenges of maintaining data quality in high-volume environments. We examine the limitations of traditional methods in managing the scale, velocity, and variety of big data and propose a conceptual approach leveraging advanced machine learning techniques. Our framework outlines a system architecture that incorporates anomaly detection, classification, and predictive analytics for real-time, scalable data quality management. Key components include an intelligent data ingestion layer, adaptive preprocessing mechanisms, context-aware feature extraction, and AI-based quality assessment modules. A continuous learning paradigm is central to our framework, ensuring adaptability to evolving data patterns and quality requirements. We also address implications for scalability, privacy, and integration within existing data ecosystems. While practical results are not provided, it lays a robust theoretical foundation for future research and implementations, advancing data quality management and encouraging the exploration of AI-driven solutions in dynamic environments.

Neural scaling laws describe how the performance of deep neural networks scales with key factors such as training data size, model complexity, and training time, often following power-law behaviors over multiple orders of magnitude. Despite their empirical observation, the theoretical understanding of these scaling laws remains limited. In this work, we employ techniques from statistical mechanics to analyze one-pass stochastic gradient descent within a student-teacher framework, where both the student and teacher are two-layer neural networks. Our study primarily focuses on the generalization error and its behavior in response to data covariance matrices that exhibit power-law spectra. For linear activation functions, we derive analytical expressions for the generalization error, exploring different learning regimes and identifying conditions under which power-law scaling emerges. Additionally, we extend our analysis to non-linear activation functions in the feature learning regime, investigating how power-law spectra in the data covariance matrix impact learning dynamics. Importantly, we find that the length of the symmetric plateau depends on the number of distinct eigenvalues of the data covariance matrix and the number of hidden units, demonstrating how these plateaus behave under various configurations. In addition, our results reveal a transition from exponential to power-law convergence in the specialized phase when the data covariance matrix possesses a power-law spectrum. This work contributes to the theoretical understanding of neural scaling laws and provides insights into optimizing learning performance in practical scenarios involving complex data structures.

Traditional generalization results in statistical learning require a training data set made of independently drawn examples. Most of the recent efforts to relax this independence assumption have considered either purely temporal (mixing) dependencies, or graph-dependencies, where non-adjacent vertices correspond to independent random variables. Both approaches have their own limitations, the former requiring a temporal ordered structure, and the latter lacking a way to quantify the strength of inter-dependencies. In this work, we bridge these two lines of work by proposing a framework where dependencies decay with graph distance. We derive generalization bounds leveraging the online-to-PAC framework, by deriving a concentration result and introducing an online learning framework incorporating the graph structure. The resulting high-probability generalization guarantees depend on both the mixing rate and the graph's chromatic number.

A common objective in statistical learning is uncertainty quantification. This task is well-studied and forms a key component of supervised and unsupervised statistical learning. In the context of regression this amounts to estimating the generalisation error, i.e., the expected mean-squared prediction error (MSPE) for unknown test observations. Usually either analytic methods for a specific model, or computational algorithms, e.g., the bootstrap or cross-validation, are employed for obtaining generalisation error estimates; see, e.g., Stone (1974), Efron & Tibshirani (1994), Chapter 7 in Hastie et al. (2017), Hastie et al. (2015), Nadeau & Bengio (1999), Wainwright (2019), Wasserman (2010). Moreover, these generalisation error estimates can be used for comparing different models, cross-validation during training, and selecting a final model from a finite collection of candidate models; see, e.g., Wood (2017), David J. Olive & Haile (2022), Das & Nason (2016).

Classical results in statistical learning typically consider two extreme data-generating models: i.i.d. instances from an unknown distribution, or fully adversarial instances, often much more challenging statistically. To bridge the gap between these models, recent work introduced the smoothed framework, in which at each iteration an adversary generates instances from a distribution constrained to have density bounded by $\sigma^{-1}$ compared to some fixed base measure $\mu$. This framework interpolates between the i.i.d. and adversarial cases, depending on the value of $\sigma$. For the classical online prediction problem, most prior results in smoothed online learning rely on the arguably strong assumption that the base measure $\mu$ is known to the learner, contrasting with standard settings in the PAC learning or consistency literature. We consider the general agnostic problem in which the base measure is unknown and values are arbitrary. Along this direction, Block et al. showed that empirical risk minimization has sublinear regret under the well-specified assumption. We propose an algorithm R-Cover based on recursive coverings which is the first to guarantee sublinear regret for agnostic smoothed online learning without prior knowledge of $\mu$. For classification, we prove that R-Cover has adaptive regret $\tilde O(\sqrt{dT/\sigma})$ for function classes with VC dimension $d$, which is optimal up to logarithmic factors. For regression, we establish that R-Cover has sublinear oblivious regret for function classes with polynomial fat-shattering dimension growth.

We consider learning under the constraint of local differential privacy (LDP). For many learning problems known efficient algorithms in this model require many rounds of communication between the server and the clients holding the data points. Yet multi-round protocols are prohibitively slow in practice due to network latency and, as a result, currently deployed large-scale systems are limited to a single round. Despite significant research interest, very little is known about which learning problems can be solved by such non-interactive systems. The only lower bound we are aware of is for PAC learning an artificial class of functions with respect to a uniform distribution (Kasiviswanathan et al., 2008).