We show a principled way of deriving online learning algorithms from a minimax analysis. Various upper bounds on the minimax value, previously thought to be non-constructive, are shown to yield algorithms. This allows us to seamlessly recover known methods and to derive new ones, also capturing such ''unorthodox'' methods as Follow the Perturbed Leader and the R 2 forecaster. Understanding the inherent complexity of the learning problem thus leads to the development of algorithms. To illustrate our approach, we present several new algorithms, including a family of randomized methods that use the idea of a ''random play out''.

We consider sequential prediction algorithms that are given the predictions from a set of models as inputs. If the nature of the data is changing over time in that different models predict well on different segments of the data, then adaptivity is typically achieved by mixing into the weights in each round a bit of the initial prior (kind of like a weak restart). However, what if the favored models in each segment are from a small subset, i.e. the data is likely to be predicted well by models that predicted well before? Curiously, fitting such ''sparse composite models'' is achieved by mixing in a bit of all the past posteriors. This self-referential updating method is rather peculiar, but it is efficient and gives superior performance on many natural data sets.

We study consistency properties of surrogate loss functions for general multiclass classification problems, defined by a general loss matrix. We extend the notion of classification calibration, which has been studied for binary and multiclass 0-1 classification problems (and for certain other specific learning problems), to the general multiclass setting, and derive necessary and sufficient conditions for a surrogate loss to be classification calibrated with respect to a loss matrix in this setting. We then introduce the notion of \emph{classification calibration dimension} of a multiclass loss matrix, which measures the smallest size' of a prediction space for which it is possible to design a convex surrogate that is classification calibrated with respect to the loss matrix. We derive both upper and lower bounds on this quantity, and use these results to analyze various loss matrices. In particular, as one application, we provide a different route from the recent result of Duchi et al.\ (2010) for analyzing the difficulty of designinglow-dimensional' convex surrogates that are consistent with respect to pairwise subset ranking losses.

The accurate prediction of molecular energetics in chemical compound space is a crucial ingredient for rational compound design. The inherently graph-like, non-vectorial nature of molecular data gives rise to a unique and difficult machine learning problem. In this paper, we adopt a learning-from-scratch approach where quantum-mechanical molecular energies are predicted directly from the raw molecular geometry. The study suggests a benefit from setting flexible priors and enforcing invariance stochastically rather than structurally. Our results improve the state-of-the-art by a factor of almost three, bringing statistical methods one step closer to the holy grail of ''chemical accuracy''.

Psychologists are interested in developing instructional policies that boost student learning. An instructional policy specifies the manner and content of instruction. For example, in the domain of concept learning, a policy might specify the nature of exemplars chosen over a training sequence. Traditional psychological studies compare several hand-selected policies, e.g., contrasting a policy that selects only difficult-to-classify exemplars with a policy that gradually progresses over the training sequence from easy exemplars to more difficult (known as {\em fading}). We propose an alternative to the traditional methodology in which we define a parameterized space of policies and search this space to identify the optimum policy.

This course introduces participants to MLOps tools and best practices for deploying, evaluating, monitoring and operating production ML systems on Google Cloud. MLOps is a discipline focused on the deployment, testing, monitoring, and automation of ML systems in production. Machine Learning Engineering professionals use tools for continuous improvement and evaluation of deployed models. They work with (or can be) Data Scientists, who develop models, to enable velocity and rigor in deploying the best performing models. This course is primarily intended for the following participants: Data Scientists looking to quickly go from machine learning prototype to production to deliver business impact.

The AI bot, known for providing restaurant recommendations and answering privacy policy questions, has now managed to concoct a sexual harassment scandal involving a real-life law professor., Technology & Science News, Times Now

Berkeley experts in artificial intelligence are studying how things like ChatGPT will transform everything from admissions screening and research to writing college essays. It passed the bar exam, first with a mediocre score and then with a ranking among the top tier of newly minted lawyers. It scored better than 90% of SAT takers. It nearly aced the verbal section of the GRE -- though it has room for improvement with AP Composition. In the months since the machine-learning interface ChatGPT debuted, hundreds of headlines and hot-takes have whirled about how artificial intelligence will overhaul everything from health care and business to legal affairs and shopping.

ABSTRACT: In this study, we have shown autonomous long-term prediction with a spintronic physical reservoir. Due to the short-term memory property of the magnetization dynamics, non-linearity arises in the reservoir states which could be used for long-term prediction tasks using simple linear regression for online training. During the prediction stage, the output is directly fed to the input of the reservoir for autonomous prediction. We employ our proposed reservoir for the modeling of the chaotic time series such as Mackey-Glass and dynamic time-series data, such as household building energy loads. Since only the last layer of a RC needs to be trained with linear regression, it is well suited for learning in real time on edge devices. Here we show that a skyrmion based magnetic tunnel junction can potentially be used as a prototypical RC but any nanomagnetic magnetic tunnel junction with nonlinear magnetization behavior can implement such a RC. By comparing our spintronic physical RC approach with state-of-the-art energy load forecasting algorithms, such as LSTMs and RNNs, we conclude that the proposed framework presents good performance in achieving high predictions accuracy, while also requiring low memory and energy both of which are at a premium in hardware resource and power constrained edge applications. Further, the proposed approach is shown to require very small training datasets and at the same time being at least 16X energy efficient compared to the state-of-the-art sequence to sequence LSTM for accurate household load predictions. I. INTRODUCTION Recurrent neural networks (RNNs) [1,2] are shown to be more suitable in temporal data processing tasks than the traditional feedforward neural networks (FNNs) because of the recurrent connections among constituent neurons. However, RNNs often suffers from vanishing and exploding gradients problem due to the long-term dependencies that could arise in the recurrent layers. To circumvent these issues variations of RNN is proposed, i.e., long short-term memory (LSTM) [3] and reservoir computing (RC) [4,5].

Automated machine learning (AutoML) methods improve upon existing models by optimizing various aspects of their design. While present methods focus on hyperparameters and neural network topologies, other aspects of neural network design can be optimized as well. To further the state of the art in AutoML, this dissertation introduces techniques for discovering more powerful activation functions and establishing more robust weight initialization for neural networks. These contributions improve performance, but also provide new perspectives on neural network optimization. First, the dissertation demonstrates that discovering solutions specialized to specific architectures and tasks gives better performance than reusing general approaches. Second, it shows that jointly optimizing different components of neural networks is synergistic, and results in better performance than optimizing individual components alone. Third, it demonstrates that learned representations are easier to optimize than hard-coded ones, creating further opportunities for AutoML. The dissertation thus makes concrete progress towards fully automatic machine learning in the future.