Adaptive gradient methods are the method of choice for optimization in machine learning and used to train the largest deep models. In this paper we study the problem of learning a local preconditioner, that can change as the data is changing along the optimization trajectory. We propose an adaptive gradient method that has provable adaptive regret guarantees vs. the best local preconditioner. To derive this guarantee, we prove a new adaptive regret bound in online learning that improves upon previous adaptive online learning methods. We demonstrate the robustness of our method in automatically choosing the optimal learning rate schedule for popular benchmarking tasks in vision and language domains. Without the need to manually tune a learning rate schedule, our method can, in a single run, achieve comparable and stable task accuracy as a fine-tuned optimizer.

Reducing reinforcement learning to supervised learning is a well-studied and effective approach that leverages the benefits of compact function approximation to deal with large-scale Markov decision processes. Independently, the boosting methodology (e.g. AdaBoost) has proven to be indispensable in designing efficient and accurate classification algorithms by combining inaccurate rules-of-thumb. In this paper, we take a further step: we reduce reinforcement learning to a sequence of weak learning problems. Since weak learners perform only marginally better than random guesses, such subroutines constitute a weaker assumption than the availability of an accurate supervised learning oracle. We prove that the sample complexity and running time bounds of the proposed method do not explicitly depend on the number of states. While existing results on boosting operate on convex losses, the value function over policies is non-convex. We show how to use a non-convex variant of the Frank-Wolfe method for boosting, that additionally improves upon the known sample complexity and running time even for reductions to supervised learning.

Conversational information seeking (CIS) is concerned with a sequence of interactions between one or more users and an information system. Interactions in CIS are primarily based on natural language dialogue, while they may include other types of interactions, such as click, touch, and body gestures. This monograph provides a thorough overview of CIS definitions, applications, interactions, interfaces, design, implementation, and evaluation. This monograph views CIS applications as including conversational search, conversational question answering, and conversational recommendation. Our aim is to provide an overview of past research related to CIS, introduce the current state-of-the-art in CIS, highlight the challenges still being faced in the community. and suggest future directions.

We prove linear convergence of gradient descent to a global optimum for the training of deep residual networks with constant layer width and smooth activation function. We show that if the trained weights, as a function of the layer index, admit a scaling limit as the depth increases, then the limit has finite $p-$variation with $p=2$. Proofs are based on non-asymptotic estimates for the loss function and for norms of the network weights along the gradient descent path. We illustrate the relevance of our theoretical results to practical settings using detailed numerical experiments on supervised learning problems.

Backward compatibility of model predictions is a desired property when updating a machine learning driven application. It allows to seamlessly improve the underlying model without introducing regression bugs. In classification tasks these bugs occur in the form of negative flips. This means an instance that was correctly classified by the old model is now classified incorrectly by the updated model. This has direct negative impact on the user experience of such systems e.g. a frequently used voice assistant query is suddenly misclassified. A common reason to update the model is when new training data becomes available and needs to be incorporated. Simply retraining the model with the updated data introduces the unwanted negative flips. We study the problem of regression during data updates and propose Backward Compatible Weight Interpolation (BCWI). This method interpolates between the weights of the old and new model and we show in extensive experiments that it reduces negative flips without sacrificing the improved accuracy of the new model. BCWI is straight forward to implement and does not increase inference cost. We also explore the use of importance weighting during interpolation and averaging the weights of multiple new models in order to further reduce negative flips.

Over the last few months, his manager began noticing a decided improvement. Timothy's reports were more detailed, more persuasive, and more original. They were also written entirely by a chatbot. Since its launch late last year, ChatGPT, an artificial-intelligence bot that can create original content that's all but indistinguishable from that of a human being, has raised thorny ethical questions in numerous quarters--from school districts, where educators worry about cheating, to the political arena, where experts have raised concerns about automated lobbying. The business world hasn't been a focus of these debates, but its leaders are also uneasy about work product created by AI.

AI text generator ChatGPT, released to the public late last year, is so sophisticated that it has already demonstrated its ability to write coherent essays, generate sound legal documents and otherwise interact with humans in a convincingly conversational manner. One CEO even treats the tool from parent company OpenAI like a perennially available member of his executive team. "I ask ChatGPT to become aware of where my biases and blindspots might be, and the answers it gives are a really, really good starting point to check your thinking," Jeff Maggioncalda, CEO of online course provider Coursera, told CBS MoneyWatch. He said the tool helps him to be more thoughtful in his approach to business challenges, as well as look at topics from vantage points that differ from his own. For example, last week at the World Economic Forum meeting in Davos, Switzerland, Maggioncalda entered the following prompt: "What should I consider when giving a speech to prime ministers at Davos?" Another useful entry for business leaders would be: "What should I consider when I am restructuring my company?"

ChatGPT, which the company OpenAI recently released, generates text and can even write essays. The artificial intelligence, or AI, chatbot has already been reported to be a coauthor on four papers and preprints. What does this mean for the future of education? How can universities best address the challenges ChatGPT, or "Chat Generative Pre-trained Transformer," poses? Could the bot enhance education?

AI has been utilized to automate jobs in a variety of businesses, and it will be useful in the educational sector as well. Professors and teachers frequently need to manage the classroom environment on top of carrying out a variety of administrative and organizational tasks. A report in research paper writing services claims that teachers do more than merely instruct. They also spend time arranging resources and materials for lectures, managing instructional materials, creating progress reports, grading tests, assessing homework, filing required paperwork and other tasks. There is a significant amount of work involved here. AI allows us to automate the administrative and management chores that institutions and instructors perform. AI assists in controlling the classroom atmosphere and numerous administrative responsibilities. The evaluation of assignments, exam grades, and many other things is also made simple by AI.

Welcome to "Tableau Tutorial for Beginners"! In this course, you will learn everything you need to know to get started with Tableau. We will begin by introducing you to the different types of Tableau products and how they can be used. You will then learn how to download and install Tableau Desktop, and how to import data into the software. Next, we will cover the basics of the Tableau interface, and show you how to build custom visualizations.