This paper studies mixed linear regression and give a number of results. Under various deterministic conditions, they show that given a sufficiently warm start, iterative trimmed least squares converges to the true directions quickly. Their algorithm continues to work in the presence of adversarial corruptions. However the warm start is required to be quite close to the true solution. They give an SVD based initialization procedure that works in the non-noisy setting and when the examples come from a gaussian distribution.

Machine learning (ML) models can be effective for forecasting the dynamics of unknown systems from time-series data, but they often require large amounts of data and struggle to generalize across systems with varying dynamics. Combined, these issues make forecasting from short time series particularly challenging. To address this problem, we introduce Meta-learning for Tailored Forecasting from Related Time Series (METAFORS), which uses related systems with longer time-series data to supplement limited data from the system of interest. By leveraging a library of models trained on related systems, METAFORS builds tailored models to forecast system evolution with limited data. Using a reservoir computing implementation and testing on simulated chaotic systems, we demonstrate METAFORS' ability to predict both short-term dynamics and long-term statistics, even when test and related systems exhibit significantly different behaviors and the available data are scarce, highlighting its robustness and versatility in data-limited scenarios.

The use of connected automated vehicle (CAV) is advocated to mitigate traffic oscillations in mixed traffic flow consisting of CAVs and human driven vehicles (HDVs). This study proposes an adaptive deep Koopman predictive control framework (AdapKoopPC) for regulating mixed traffic flow. Firstly, a Koopman theory-based adaptive trajectory prediction deep network (AdapKoopnet) is designed for modeling HDVs car-following behavior. AdapKoopnet enables the representation of HDVs behavior by a linear model in a high-dimensional space. Secondly, the model predictive control is employed to smooth the mixed traffic flow, where the combination of the linear dynamic model of CAVs and linear prediction blocks from AdapKoopnet is embedded as the predictive model into the AdapKoopPC. Finally, the predictive performance of the prosed AdapKoopnet is verified using the HighD naturalistic driving dataset. Furthermore, the control performance of AdapKoopPC is validated by the numerical simulations. Results demonstrate that the AdapKoopnet provides more accuracy HDVs predicted trajectories than the baseline nonlinear models. Moreover, the proposed AdapKoopPC exhibits more effective control performance with less computation cost compared with baselines in mitigating traffic oscillations, especially at the low CAVs penetration rates. The code of proposed AdapKoopPC is open source.

State-of-the-art methods for backpropagation-free learning employ local error feedback to direct iterative optimisation via gradient descent. In this study, we examine the more restrictive setting where retrograde communication from neuronal outputs is unavailable for pre-synaptic weight optimisation. To address this challenge, we propose Forward Projection (FP). This novel randomised closed-form training method requires only a single forward pass over the entire dataset for model fitting, without retrograde communication. Target values for pre-activation membrane potentials are generated layer-wise via nonlinear projections of pre-synaptic inputs and the labels. Local loss functions are optimised over pre-synaptic inputs using closed-form regression, without feedback from neuronal outputs or downstream layers. Interpretability is a key advantage of FP training; membrane potentials of hidden neurons in FP-trained networks encode information which is interpretable layer-wise as label predictions. We demonstrate the effectiveness of FP across four biomedical datasets. In few-shot learning tasks, FP yielded more generalisable models than those optimised via backpropagation. In large-sample tasks, FP-based models achieve generalisation comparable to gradient descent-based local learning methods while requiring only a single forward propagation step, achieving significant speed up for training. Interpretation functions defined on local neuronal activity in FP-based models successfully identified clinically salient features for diagnosis in two biomedical datasets. Forward Projection is a computationally efficient machine learning approach that yields interpretable neural network models without retrograde communication of neuronal activity during training.

The momentum Schr\"odinger Bridge (mSB) has emerged as a leading method for accelerating generative diffusion processes and reducing transport costs. However, the lack of simulation-free properties inevitably results in high training costs and affects scalability. To obtain a trade-off between transport properties and scalability, we introduce variational Schr\"odinger momentum diffusion (VSMD), which employs linearized forward score functions (variational scores) to eliminate the dependence on simulated forward trajectories. Our approach leverages a multivariate diffusion process with adaptively transport-optimized variational scores. Additionally, we apply a critical-damping transform to stabilize training by removing the need for score estimations for both velocity and samples. Theoretically, we prove the convergence of samples generated with optimal variational scores and momentum diffusion. Empirical results demonstrate that VSMD efficiently generates anisotropic shapes while maintaining transport efficacy, outperforming overdamped alternatives, and avoiding complex denoising processes. Our approach also scales effectively to real-world data, achieving competitive results in time series and image generation.

Asynchronous Stochastic Gradient Descent (Asynchronous SGD) is a cornerstone method for parallelizing learning in distributed machine learning. However, its performance suffers under arbitrarily heterogeneous computation times across workers, leading to suboptimal time complexity and inefficiency as the number of workers scales. While several Asynchronous SGD variants have been proposed, recent findings by Tyurin & Richt\'arik (NeurIPS 2023) reveal that none achieve optimal time complexity, leaving a significant gap in the literature. In this paper, we propose Ringmaster ASGD, a novel Asynchronous SGD method designed to address these limitations and tame the inherent challenges of Asynchronous SGD. We establish, through rigorous theoretical analysis, that Ringmaster ASGD achieves optimal time complexity under arbitrarily heterogeneous and dynamically fluctuating worker computation times. This makes it the first Asynchronous SGD method to meet the theoretical lower bounds for time complexity in such scenarios.

Large language models have achieved remarkable success in recent years, primarily due to the implementation of self-attention mechanisms. However, traditional Softmax attention suffers from numerical instability and reduced performance as the length of inference tokens increases. This paper addresses these issues by decomposing the Softmax operation into a non-linear transformation and the $l_1$-norm. We identify the latter as essential for maintaining model performance. By replacing the non-linear transformation with the Softplus activation function and introducing a dynamic scale factor for different token lengths based on invariance entropy, we create a novel attention mechanism with performance better than conventional Softmax attention across various inference lengths. To further improve the length extrapolation ability of the proposed attention mechanism, we introduce a fine-tuning-free re-weighting mechanism that amplifies significant attention weights while diminishing weaker ones, enabling the model to concentrate more effectively on relevant tokens without requiring retraining. When combined with our proposed attention mechanism, this approach demonstrates significant promise in managing longer sequences, maintaining nearly constant validation loss even at 16$\times$ the training token length while ensuring numerical stability. Our code is available at: https://github.com/iminfine/freeatten.

We present a $k$-means-based clustering algorithm, which optimizes the mean square error, for given cluster sizes. A straightforward application is balanced clustering, where the sizes of each cluster are equal. In the $k$-means assignment phase, the algorithm solves an assignment problem using the Hungarian algorithm. This makes the assignment phase time complexity $O(n^3)$. This enables clustering of datasets of size more than 5000 points.

While attention-based models have demonstrated the remarkable ability of in-context learning, the theoretical understanding of how these models acquired this ability through gradient descent training is still preliminary. Towards answering this question, we study the gradient descent dynamics of multi-head linear self-attention trained for in-context linear regression. We examine two parametrizations of linear self-attention: one with the key and query weights merged as a single matrix (common in theoretical studies), and one with separate key and query matrices (closer to practical settings). For the merged parametrization, we show the training dynamics has two fixed points and the loss trajectory exhibits a single, abrupt drop. We derive an analytical time-course solution for a certain class of datasets and initialization. For the separate parametrization, we show the training dynamics has exponentially many fixed points and the loss exhibits saddle-to-saddle dynamics, which we reduce to scalar ordinary differential equations. During training, the model implements principal component regression in context with the number of principal components increasing over training time. Overall, we characterize how in-context learning abilities evolve during gradient descent training of linear attention, revealing dynamics of abrupt acquisition versus progressive improvements in models with different parametrizations.

Copyright policies play a pivotal role in protecting the intellectual property of creators and companies in creative industries. The advent of cost-reducing technologies, such as generative AI, in these industries calls for renewed attention to the role of these policies. This paper studies product positioning and competition in a market of creatively differentiated products and the competitive and welfare effects of copyright protection. A common feature of products with creative elements is that their key attributes (e.g., images and text) are unstructured and thus high-dimensional. We focus on a stylized design product, fonts, and use data from the world's largest online marketplace for fonts. We use neural network embeddings to quantify unstructured attributes and measure the visual similarity. We show that this measure closely aligns with actual human perception. Based on this measure, we empirically find that competitions occur locally in the visual characteristics space. We then develop a structural model for supply and demand that integrate the embeddings. Through counterfactual analyses, we find that local copyright protection can enhance consumer welfare when products are relocated, and the interplay between copyright and cost-reducing technologies is essential in determining an optimal policy for social welfare. We believe that the embedding analysis and empirical models introduced in this paper can be applicable to a range of industries where unstructured data captures essential features of products and markets.