Meta-Learning

In-context learning (ICL) allows large language models to adapt to new tasks from a few examples without updating their parameters. Existing theories explain ICL by assuming the test task distribution matches pretraining -- an assumption that breaks down under adversarial distribution shifts. We introduce a distributionally robust meta-learning framework that provides worst-case guarantees for ICL under Wasserstein-based distribution shifts. Focusing on linear self-attention Transformers, we derive a non-asymptotic bound connecting adversarial perturbation strength ($ρ$), model capacity ($m$), and the number of in-context examples ($N$). The analysis reveals that the maximum safe perturbation radius scales as $ρ_{\max} \propto \sqrt{m}$, while maintaining performance under adversarial shift requires additional in-context examples with $N_ρ- N_0 \propto ρ^2$. Experiments on synthetic tasks confirm these scaling laws, and experiments on 21 real pretrained models (0.1B--7B parameters, 5 families) provide qualitative evidence consistent with the theory's predictions, while revealing that ICL capability is a prerequisite for robustness. These findings advance the theoretical understanding of ICL under adversarial conditions and formalize the sense in which larger models are safer under distributional shift.

Large language models (LLMs) often struggle to learn from corrective feedback within a conversational context. They are rarely proactive in soliciting this feedback, even when faced with ambiguity, which can make their dialogues feel static, one-sided, and lacking the adaptive qualities of human conversation. To address these limitations, we draw inspiration from social meta-learning (SML) in humans - the process of learning how to learn from others. We formulate SML as a finetuning methodology, training LLMs to solicit and learn from language feedback in simulated pedagogical dialogues, where static tasks are converted into interactive social learning problems. SML effectively teaches models to use conversation to solve problems they are unable to solve in a single turn. This capability generalises across domains; SML on math problems produces models that better use feedback to solve coding problems and vice versa. Furthermore, despite being trained only on fully-specified problems, these models are better able to solve underspecified tasks where critical information is revealed over multiple turns. When faced with this ambiguity, SML-trained models make fewer premature answer attempts and are more likely to ask for the information they need. This work presents a scalable approach to developing AI systems that effectively learn from language feedback.

Current meta-learning methods are constrained to narrow task distributions with fixed feature and label spaces, limiting applicability. Moreover, the current meta-learning literature uses key terms like "universal" and "general-purpose" inconsistently and lacks precise definitions, hindering comparability. We introduce a theoretical framework for meta-learning which formally defines practical universality and introduces a distinction between algorithm-explicit and algorithm-implicit learning, providing a principled vocabulary for reasoning about universal meta-learning methods. Guided by this framework, we present TAIL, a transformer-based algorithm-implicit meta-learner that functions across tasks with varying domains, modalities, and label configurations. TAIL features three innovations over prior transformer-based meta-learners: random projections for cross-modal feature encoding, random injection label embeddings that extrapolate to larger label spaces, and efficient inline query processing. TAIL achieves state-of-the-art performance on standard few-shot benchmarks while generalizing to unseen domains. Unlike other meta-learning methods, it also generalizes to unseen modalities, solving text classification tasks despite training exclusively on images, handles tasks with up to 20$\times$ more classes than seen during training, and provides orders-of-magnitude computational savings over prior transformer-based approaches.

Continual learning aims to acquire tasks sequentially without catastrophic forgetting, yet standard strategies face a core tradeoff: regularization-based methods (e.g., EWC) can overconstrain updates when task optima are weakly overlapping, while replay-based methods can retain performance but drift due to imperfect replay. We study a hybrid perspective: trust region continual learning that combines generative replay with a Fisher-metric trust region constraint. We show that, under local approximations, the resulting update admits a Model-Agnostic Meta-Learning (MAML) interpretation with a single implicit inner step: replay supplies an old-task gradient signal (query-like), while the Fisher-weighted penalty provides efficient offline curvature shaping (support-like). This yields an emergent meta-learning property in continual learning: the model becomes an initialization that rapidly re-converges to prior task optima after each task transition, without explicitly optimizing a bilevel objective that couples inner adaptation with outer meta-updates. Empirically, on task-incremental diffusion image generation and continual diffusion-policy control, trust region continual learning achieves the best final performance and retention, and accelerates early-task recovery compared with EWC, replay, and continual meta-learning baselines.

Distributed Fiber Optic Sensing (DFOS) has emerged as a promising technology for long-range and real-time perimeter security in critical infrastructure monitoring. However, DFOS signals collected from different field deployments often exhibit substantial distribution shifts caused by variations in fiber installation, structural coupling, and environmental noise. These deployment-dependent changes make reliable event recognition difficult in practical perimeter security systems, especially when labeled samples from new target sites are scarce or unavailable. To address these challenges, this paper proposes DUPLE, an intelligent cross-deployment recognition framework for fiber-optic perimeter security under label-scarce target deployments. DUPLE employs statistically guided meta-learning to enhance recognition robustness across unseen deployments. Specifically, a dual-domain multi-prototype learner jointly models temporal and frequency-domain evidence to capture intra-class variability under deployment shifts. A statistical guidance network estimates sample-specific domain reliability from raw signal statistics, while a query-aware aggregation mechanism adaptively selects relevant prototypes for each test sample. Extensive experiments on two real-world cross-deployment DFOS benchmarks demonstrate that DUPLE consistently outperforms representative traditional machine learning, deep learning, domain generalization, and meta-learning baselines. Ablation, few-shot, per-deployment, and efficiency analyses further verify the effectiveness and practicality of DUPLE for reliable DFOS-based perimeter security monitoring.

Decision trees are widely used in high-stakes fields like finance and healthcare due to their interpretability. This work introduces an efficient, scalable method for generating synthetic pre-training data to enable meta-learning of decision trees. Our approach samples near-optimal decision trees synthetically, creating large-scale, realistic datasets. Using the MetaTree transformer architecture, we demonstrate that this method achieves performance comparable to pre-training on real-world data or with computationally expensive optimal decision trees. This strategy significantly reduces computational costs, enhances data generation flexibility, and paves the way for scalable and efficient meta-learning of interpretable decision tree models.

Uncovering the causal mechanisms of complex real-world systems remains a significant challenge, as these systems often entail high data collection costs and involve unknown interventions. We introduce MetaCaDI, the first framework to cast the identification of unknown interventions as a meta-learning problem, explicitly leveraging a jointly learned causal graph. MetaCaDI is a Bayesian framework that learns a shared causal structure across multiple environments and is optimized to rapidly adapt to new, few-shot intervention target identification tasks. A key innovation is our model's analytical adaptation, which uses a closed-form solution to bypass expensive and potentially unstable gradient-based bilevel optimization. Extensive experiments on synthetic and complex gene expression data demonstrate that MetaCaDI significantly outperforms state-of-the-art methods. It excels at identifying intervention targets from as few as 3 samples - where existing methods collapse to random chance - while robustly recovering the shared causal graph, proving its effectiveness in data-scarce scenarios.

Dynamic Voltage and Frequency Scaling is essential for enhancing energy efficiency in mobile platforms. However, traditional heuristic-based governors are increasingly inadequate for managing the complexity of heterogeneous System-on-Chip designs and diverse application workloads. Although reinforcement learning approaches offer improved performance, their poor generalization capability and reliance on extensive retraining for each hardware and application combination leads to significant deployment costs. In this work, we observe that device and application metadata inherently encapsulate valuable knowledge for DVFS, presenting an opportunity to overcome these limitations. We formulate DVFS for heterogeneous devices and applications as a multi-task reinforcement learning problem. We introduce MetaDVFS, which is a metadata-guided framework that systematically leverages metadata to discover and transfer shared knowledge across DVFS tasks. MetaDVFS can output a set of DVFS models with significant generalization capability for various applications of heterogeneous devices. Evaluations on five Google Pixel devices running six applications show that MetaDVFS achieves up to 17% improvement in Performance-Power Ratio and up to 26% improvement in Quality of Experience. Compared to state-of-the-art methods, MetaDVFS delivers 70.8% faster adaptation and 5.8-27.6% higher performance over standalone device-application specific training, while avoiding negative transfer effects. These results establish MetaDVFS as an effective and scalable solution for DVFS deployment in heterogeneous mobile environments.

Meta-learning aims to leverage information across related tasks to improve prediction on unlabeled data for new tasks when only a small number of labeled observations are available ("few-shot" learning). Increased task diversity is often believed to enhance meta-learning by providing richer information across tasks. However, recent work by Kumar et al. (2022) shows that increasing task diversity, quantified through the overall geometric spread of task representations, can in fact degrade meta-learning prediction performance across a range of models and datasets. In this work, we build on this observation by showing that meta-learning performance is affected not only by the overall geometric variability of task parameters, but also by how this variability is allocated relative to an underlying low-dimensional structure. Similar to Pimonova et al. (2025), we decompose task-specific regression effects into a structurally informative component and an orthogonal, non-informative component. We show theoretically and through simulation that meta-learning prediction degrades when a larger fraction of between-task variability lies in orthogonal, non-informative directions, even when the overall geometric variability of tasks is held fixed.

Human cognition is profoundly shaped by the environments in which it unfolds. Yet, it remains an open question whether learning and decision making can be explained as a principled adaptation to the statistical structure of real-world tasks. We introduce ecologically rational analysis, a computational framework that unifies the normative foundations of rational analysis with ecological grounding. Leveraging large language models to generate ecologically valid cognitive tasks at scale, and using meta-learning to derive rational models optimized for these environments, we develop a new class of learning algorithms: Ecologically Rational Meta-learned Inference (ERMI). ERMI internalizes the statistical regularities of naturalistic problem spaces and adapts flexibly to novel situations, without requiring hand-crafted heuristics or explicit parameter updates. We show that ERMI captures human behavior across 15 experiments spanning function learning, category learning, and decision making, outperforming several established cognitive models in trial-by-trial prediction. Our results suggest that much of human cognition may reflect adaptive alignment to the ecological structure of the problems we encounter in everyday life.

The preparation of quantum Gibbs states is a fundamental challenge in quantum computing, essential for applications ranging from modeling open quantum systems to quantum machine learning. Building on the Meta-Variational Quantum Eigensolver framework proposed by Cervera-Lierta et al. and a problem-driven ansatz design, we introduce two meta-learning algorithms: Meta-Variational Quantum Thermalizer (Meta-VQT) and Neural Network Meta-VQT (NN-Meta VQT) for efficient thermal state preparation of parametrized Hamiltonians on Noisy Intermediate-Scale Quantum (NISQ) devices. Meta-VQT utilizes a fully quantum ansatz, while NN Meta-VQT integrates a quantum--classical hybrid architecture. Both use collective optimization over a finite set of training parameter values to prepare approximate Gibbs states at held-out parameter values within the sampled parameter domain. We validate our methods on Transverse Field Ising Models (TFIM) with up to eight qubits and on a two-qubit Heisenberg model with all field terms. Within these finite-size benchmarks, the trained models accurately prepare thermal states at test points inside the training interval. For the six and eight qubit cases of TFIM, meta-learned parameters when combined with appropriately designed ansatz serve as warm-start initializations, significantly outperforming random initializations in the optimization tasks. Furthermore, a 3 qubit Kitaev ring example showcases our algorithm's effectiveness across finite-temperature crossover regimes. Finally, we apply our algorithms to train a Quantum Boltzmann Machine (QBM) on a 2-qubit Heisenberg model with all field terms, achieving enhanced training efficiency, improved Gibbs state accuracy, and a 30-fold runtime speedup over existing techniques such as variational quantum imaginary time (VarQITE)-based QBM highlighting the practicality of meta-algorithm-based QBMs.

System identification has greatly benefited from deep learning techniques, particularly for modeling complex, nonlinear dynamical systems with partially unknown physics where traditional approaches may not be feasible. However, deep learning models often require large datasets and significant computational resources for both training and inference due to their high-dimensional parameterizations. To address this challenge, we propose a meta-learning framework that discovers a low-dimensional manifold within the parameter space of an over-parameterized neural network architecture. This manifold is learned from an abundant meta-dataset of input-output sequences generated by a simulator of the class of systems of interest, and thereafter enables efficient model training while preserving the network's expressive power for the considered system class. Unlike bilevel meta-learning approaches, our method employs an auxiliary neural network to map datasets directly onto the learned manifold, eliminating the need for costly second-order gradient computations during meta-training and generally reducing the number of first-order updates required in inference, which could be expensive for large models. Finally, we derive a probabilistic bound that quantifies the performance gap between training on the learned low-dimensional manifold and the full parameterization. We validate our approach on a family of Bouc-Wen oscillators, which is a well-studied nonlinear system identification benchmark. We demonstrate that we are able to learn accurate models even in small-data scenarios.

Meta-learning represents a strong class of approaches for solving few-shot learning tasks. Nonetheless, recent research suggests that simply pre-training a generic encoder can potentially surpass meta-learning algorithms. In this paper, we first discuss the reasons why meta-learning fails to stand out in these few-shot learning experiments, and hypothesize that it is due to the few-shot learning tasks lacking diversity. We propose DRESS, a task-agnostic Disentangled REpresentation-based Self-Supervised meta-learning approach that enables fast model adaptation on highly diversified few-shot learning tasks. Specifically, DRESS utilizes disentangled representation learning to create self-supervised tasks that can fuel the meta-training process. Furthermore, we also propose a class-partition based metric for quantifying the task diversity directly on the input space. We validate the effectiveness of DRESS through experiments on datasets with multiple factors of variation and varying complexity. The results suggest that DRESS is able to outperform competing methods on the majority of the datasets and task setups. Through this paper, we advocate for a re-examination of proper setups for task adaptation studies, and aim to reignite interest in the potential of meta-learning for solving few-shot learning tasks via disentangled representations.

Pairwise difference learning (PDL) has recently been introduced as a new meta-learning technique for regression. Instead of learning a mapping from instances to outcomes in the standard way, the key idea is to learn a function that takes two instances as input and predicts the difference between the respective outcomes. Given a function of this kind, predictions for a query instance are derived from every training example and then averaged. This paper extends PDL toward the task of classification and proposes a meta-learning technique for inducing a PDL classifier by solving a suitably defined (binary) classification problem on a paired version of the original training data. We analyze the performance of the PDL classifier in a large-scale empirical study and find that it outperforms state-of-the-art methods in terms of prediction performance. Last but not least, we provide an easy-to-use and publicly available implementation of PDL in a Python package.

Is it possible to understand the intricacies of a dynamical system not solely from its input/output pattern, but also by observing the behavior of other systems within the same class? This central question drives the study presented in this paper. In response to this query, we introduce a novel paradigm for system identification, addressing two primary tasks: one-step-ahead prediction and multi-step simulation. Unlike conventional methods, we do not directly estimate a model for the specific system. Instead, we learn a meta model that represents a class of dynamical systems. This meta model is trained on a potentially infinite stream of synthetic data, generated by simulators whose settings are randomly extracted from a probability distribution. When provided with a context from a new system-specifically, an input/output sequence-the meta model implicitly discerns its dynamics, enabling predictions of its behavior. The proposed approach harnesses the power of Transformers, renowned for their in-context learning capabilities. For one-step prediction, a GPT-like decoder-only architecture is utilized, whereas the simulation problem employs an encoder-decoder structure. Initial experimental results affirmatively answer our foundational question, opening doors to fresh research avenues in system identification.

Selecting an appropriate classification algorithm for a given data set remains a challenging problem in data mining and machine learning. Existing algorithm recommendation models are typically trained with individual learners and rely on only one type of meta-feature, which may limit their ability to capture the diverse characteristics of classification problems. This paper proposes a multi-view ensemble meta-learning framework for classification algorithm recommendation. The framework constructs base recommendation models from different combinations of heterogeneous meta-feature groups and combines them through an accuracy- and diversity-aware ensemble strategy. The main focus of this work is empirical: we evaluate the proposed method on 1,090 benchmark classification problems derived from 84 public data sets, using 13 widely used candidate classification algorithms and five types of meta-features. The experimental results show that the proposed ensemble recommendation method consistently improves ranking loss, average precision, and top-ranked recommendation precision over individual recommendation models. These results suggest that combining complementary meta-feature views is an effective strategy for robust classification algorithm recommendation.