Theo Wolf

Alexander D. Goldie, Zilin Wang, Adrian Hayler et al.

Automating the development of machine learning algorithms has the potential to unlock new breakthroughs. However, our ability to improve and evaluate algorithm discovery systems has thus far been limited by existing task suites. They suffer from many issues, such as: poor evaluation methodologies; data contamination; and containing saturated or very similar problems. Here, we introduce DiscoGen, a procedural generator of algorithm discovery tasks for machine learning, such as developing optimisers for reinforcement learning or loss functions for image classification. Motivated by the success of procedural generation in reinforcement learning, DiscoGen spans millions of tasks of varying difficulty and complexity from a range of machine learning fields. These tasks are specified by a small number of configuration parameters and can be used to optimise algorithm discovery agents (ADAs). We present DiscoBench, a benchmark consisting of a fixed, small subset of DiscoGen tasks for principled evaluation of ADAs. Finally, we propose a number of ambitious, impactful research directions enabled by DiscoGen, in addition to experiments demonstrating its use for prompt optimisation of an ADA. DiscoGen is released open-source at https://github.com/AlexGoldie/discogen.

Maria Conchita Agana Navarro, Geng Li, Theo Wolf et al.

The accelerating pace of climate change introduces profound non-stationarities that challenge the ability of Machine Learning based climate emulators to generalize beyond their training distributions. While these emulators offer computationally efficient alternatives to traditional Earth System Models, their reliability remains a potential bottleneck under "no-analog" future climate states, which we define here as regimes where external forcing drives the system into conditions outside the empirical range of the historical training data. A fundamental challenge in evaluating this reliability is data contamination; because many models are trained on simulations that already encompass future scenarios, true out-of-distribution (OOD) performance is often masked. To address this, we benchmark the OOD robustness of three state-of-the-art architectures: U-Net, ConvLSTM, and the ClimaX foundation model specifically restricted to a historical-only training regime (1850-2014). We evaluate these models using two complementary strategies: (i) temporal extrapolation to the recent climate (2015-2023) and (ii) cross-scenario forcing shifts across divergent emission pathways. Our analysis within this experimental setup reveals an accuracy vs. stability trade-off: while the ClimaX foundation model achieves the lowest absolute error, it exhibits higher relative performance changes under distribution shifts, with precipitation errors increasing by up to 8.44% under extreme forcing scenarios. These findings suggest that when restricted to historical training dynamics, even high-capacity foundation models are sensitive to external forcing trajectories. Our results underscore the necessity of scenario-aware training and rigorous OOD evaluation protocols to ensure the robustness of climate emulators under a changing climate.

Bidipta Sarkar, Mattie Fellows, Juan Agustin Duque et al.

We introduce Evolution Guided General Optimization via Low-rank Learning (EGGROLL), an evolution strategies (ES) algorithm designed to scale backprop-free optimization to large population sizes for modern large neural network architectures with billions of parameters. ES is a set of powerful blackbox optimisation methods that can handle non-differentiable or noisy objectives with excellent scaling potential through parallelisation. Na{ï}ve ES becomes prohibitively expensive at scale due to the computational and memory costs associated with generating matrix perturbations $E\in\mathbb{R}^{m\times n}$ and the batched matrix multiplications needed to compute per-member forward passes. EGGROLL overcomes these bottlenecks by generating random matrices $A\in \mathbb{R}^{m\times r},\ B\in \mathbb{R}^{n\times r}$ with $r\ll \min(m,n)$ to form a low-rank matrix perturbation $A B^\top$ that are used in place of the full-rank perturbation $E$. As the overall update is an average across a population of $N$ workers, this still results in a high-rank update but with significant memory and computation savings, reducing the auxiliary storage from $mn$ to $r(m+n)$ per layer and the cost of a forward pass from $\mathcal{O}(mn)$ to $\mathcal{O}(r(m+n))$ when compared to full-rank ES. A theoretical analysis reveals our low-rank update converges to the full-rank update at a fast $\mathcal{O}\left(\frac{1}{r}\right)$ rate. Our experiments show that (1) EGGROLL does not compromise the performance of ES in tabula-rasa RL settings, despite being faster, (2) it is competitive with GRPO as a technique for improving LLM reasoning, and (3) EGGROLL enables stable pre-training of nonlinear recurrent language models that operate purely in integer datatypes.

Maria Conchita Agana Navarro, Geng Li, Theo Wolf et al.

Climate change is accelerating the frequency and severity of unprecedented events, deviating from established patterns. Predicting these out-of-distribution (OOD) events is critical for assessing risks and guiding climate adaptation. While machine learning (ML) models have shown promise in providing precise, high-speed climate predictions, their ability to generalize under distribution shifts remains a significant limitation that has been underexplored in climate contexts. This research systematically evaluates state-of-the-art ML-based climate models in diverse OOD scenarios by adapting established OOD evaluation methodologies to climate data. Experiments on large-scale datasets reveal notable performance variability across scenarios, shedding light on the strengths and limitations of current models. These findings underscore the importance of robust evaluation frameworks and provide actionable insights to guide the reliable application of ML for climate risk forecasting.