Test-time Scaling (TTS) has been demonstrated to significantly enhance the reasoning capabilities of Large Language Models (LLMs) during the inference phase without altering model parameters. However, existing TTS methods are largely independent, implying that LLMs have not yet evolved to progressively learn how to scale more effectively. With the objective of evolving LLMs to learn "how to scale test-time computation," we propose LatentEvolve, a self-evolving latent TTS framework inspired by the complementary learning system (CLS) theory. Analogous to the human brain's dual system of a fast-recall hippocampus and a slow-consolidating neocortex, LatentEvolve comprises two evolutionary components: daytime scaling, which rapidly retrieves historical latent representations to better guide current LLM reasoning; and nighttime scaling, which integrates past latent optimizations in a manner akin to the human brain's consolidation of experiences during sleep. The alternation of daytime and nighttime processes facilitates a fast and slow evolution of LLM TTS, mirroring human cognitive dynamics in a fully unsupervised manner. Extensive experiments across eight benchmarks and five model backbones demonstrate that our LatentEvolve surpasses state-of-the-art TTS methods such as LatentSeek and TTRL by up to 13.33% and exhibits exceptional cross-domain and cross-backbone generalization. Much of this success in recent years has been driven by training-time scaling, wherein increasing the volume of training data and parameters consistently yields performance improvements (Kaplan et al., 2020; Aghajanyan et al., 2023). However, the pace of this scaling, particularly in terms of pre-training scale, has begun to slow, constrained by its resource-intensive nature and the depletion of high-quality training data (Villalobos et al., 2022; Zhou et al., 2025).

Humans quickly learn new concepts from a small number of examples. Replicating this capacity with Artificial Intelligence (AI) systems has proven to be challenging. When it comes to learning subjective tasks-where there is an evident scarcity of data-this capacity needs to be recreated. In this work, we propose an intuitive human-agent teaching architecture in which the human can teach an agent how to perform a task by providing demonstrations, i.e., examples. To have an intuitive interaction, we argue that the agent should be able to learn incrementally from a few single examples. To allow for this, our objective is to broaden the agent's task understanding using domain knowledge. Then, using a learning method to enable the agent to learn efficiently from a limited number of examples. Finally, to optimize how human can select the most representative and less redundant examples to provide the agent with. We apply our proposed method to the subjective task of ingredient substitution, where the agent needs to learn how to substitute ingredients in recipes based on human examples. We replicate human input using the Recipe1MSubs dataset. In our experiments, the agent achieves half its task performance after only 100 examples are provided, compared to the complete training set of 50k examples. We show that by providing examples in strategic order along with a learning method that leverages external symbolic knowledge, the agent can generalize more efficiently.

Learning, whether natural or artificial, is a process of selection. It starts with a set of candidate options and selects the more successful ones. In the case of machine learning the selection is done based on empirical estimates of prediction accuracy of candidate prediction rules on some data. Due to randomness of data sampling the empirical estimates are inherently noisy, leading to selection under uncertainty. The book provides statistical tools to obtain theoretical guarantees on the outcome of selection under uncertainty. We start with concentration of measure inequalities, which are the main statistical instrument for controlling how much an empirical estimate of expectation of a function deviates from the true expectation. The book covers a broad range of inequalities, including Markov's, Chebyshev's, Hoeffding's, Bernstein's, Empirical Bernstein's, Unexpected Bernstein's, kl, and split-kl. We then study the classical (offline) supervised learning and provide a range of tools for deriving generalization bounds, including Occam's razor, Vapnik-Chervonenkis analysis, and PAC-Bayesian analysis. The latter is further applied to derive generalization guarantees for weighted majority votes. After covering the offline setting, we turn our attention to online learning. We present the space of online learning problems characterized by environmental feedback, environmental resistance, and structural complexity. A common performance measure in online learning is regret, which compares performance of an algorithm to performance of the best prediction rule in hindsight, out of a restricted set of prediction rules. We present tools for deriving regret bounds in stochastic and adversarial environments, and under full information and bandit feedback.

Technology-enhanced learning environments often help students retrieve relevant learning content for questions arising during self-paced study. Large language models (LLMs) have emerged as novel aids for information retrieval during learning. While LLMs are effective for general-purpose question-answering, they typically lack alignment with the domain knowledge of specific course materials such as textbooks and slides. We investigate Retrieval-Augmented Generation (RAG) and GraphRAG, a knowledge graph-enhanced RAG approach, for page-level question answering in an undergraduate mathematics textbook. While RAG has been effective for retrieving discrete, contextually relevant passages, GraphRAG may excel in modeling interconnected concepts and hierarchical knowledge structures. We curate a dataset of 477 question-answer pairs, each tied to a distinct textbook page. We then compare the standard embedding-based RAG methods to GraphRAG for evaluating both retrieval accuracy-whether the correct page is retrieved-and generated answer quality via F1 scores. Our findings show that embedding-based RAG achieves higher retrieval accuracy and better F1 scores compared to GraphRAG, which tends to retrieve excessive and sometimes irrelevant content due to its entity-based structure. We also explored re-ranking the retrieved pages with LLM and observed mixed results, including performance drop and hallucinations when dealing with larger context windows. Overall, this study highlights both the promises and challenges of page-level retrieval systems in educational contexts, emphasizing the need for more refined retrieval methods to build reliable AI tutoring solutions in providing reference page numbers.

In this paper, we offer a simple argument for the conclusion that the outputs of large language models (LLMs) are meaningless. Our argument is based on two key premises: (a) that certain kinds of intentions are needed in order for LLMs' outputs to have literal meanings, and (b) that LLMs cannot plausibly have the right kinds of intentions. We defend this argument from various types of responses, for example, the semantic externalist argument that deference can be assumed to take the place of intentions and the semantic internalist argument that meanings can be defined purely in terms of intrinsic relations between concepts, such as conceptual roles. We conclude the paper by discussing why, even if our argument is sound, the outputs of LLMs nevertheless seem meaningful and can be used to acquire true beliefs and even knowledge.

This survey examines multilingual vision-language models that process text and images across languages. We review 31 models and 21 benchmarks, spanning encoder-only and generative architectures, and identify a key tension between language neutrality (consistent cross-lingual representations) and cultural awareness (adaptation to cultural contexts). Current training methods favor neutrality through contrastive learning, while cultural awareness depends on diverse data. Two-thirds of evaluation benchmarks use translation-based approaches prioritizing semantic consistency, though recent work incorporates culturally grounded content. We find discrepancies in cross-lingual capabilities and gaps between training objectives and evaluation goals.

Many institutions are currently grappling with teaching artificial intelligence (AI) in the face of growing demand and relevance in our world. The Computing Research Association (CRA) has conducted 32 moderated virtual roundtable discussions of 202 experts committed to improving AI education. These discussions slot into four focus areas: AI Knowledge Areas and Pedagogy, Infrastructure Challenges in AI Education, Strategies to Increase Capacity in AI Education, and AI Education for All. Roundtables were organized around institution type to consider the particular goals and resources of different AI education environments. We identified the following high-level community needs to increase capacity in AI education. A significant digital divide creates major infrastructure hurdles, especially for smaller and under-resourced institutions. These challenges manifest as a shortage of faculty with AI expertise, who also face limited time for reskilling; a lack of computational infrastructure for students and faculty to develop and test AI models; and insufficient institutional technical support. Compounding these issues is the large burden associated with updating curricula and creating new programs. To address the faculty gap, accessible and continuous professional development is crucial for faculty to learn about AI and its ethical dimensions. This support is particularly needed for under-resourced institutions and must extend to faculty both within and outside of computing programs to ensure all students have access to AI education. We have compiled and organized a list of resources that our participant experts mentioned throughout this study. These resources contribute to a frequent request heard during the roundtables: a central repository of AI education resources for institutions to freely use across higher education.

Multimodal Procedural Planning (MPP) aims to generate step-by-step instructions that combine text and images, with the central challenge of preserving object-state consistency across modalities while producing informative plans. Existing approaches often leverage large language models (LLMs) to refine textual steps; however, visual object-state alignment and systematic evaluation are largely underexplored. We present MMPlanner, a zero-shot MPP framework that introduces Object State Reasoning Chain-of-Thought (OSR-CoT) prompting to explicitly model object-state transitions and generate accurate multimodal plans. To assess plan quality, we design LLM-as-a-judge protocols for planning accuracy and cross-modal alignment, and further propose a visual step-reordering task to measure temporal coherence. Experiments on RECIPEPLAN and WIKIPLAN show that MMPlanner achieves state-of-the-art performance, improving textual planning by +6.8%, cross-modal alignment by +11.9%, and visual step ordering by +26.7%