Why Experts Can't Agree on Whether AI Has a Mind Pillay is an editorial fellow at TIME. Pillay is an editorial fellow at TIME. I'm not used to getting nasty emails from a holy man, says Professor Michael Levin, a developmental biologist at Tufts University. Levin was presenting his research to a group of engineers interested in spiritual matters in India, arguing that properties like "mind" and intelligence can be observed even in cellular systems, and that they exist on a spectrum. But when he pushed further--arguing that the same properties emerge everywhere, including in computers--the reception shifted.

Artificial intelligence has advanced rapidly through large neural networks trained on massive datasets using thousands of GPUs or TPUs. Such training can occupy entire data centers for weeks and requires enormous computational and energy resources. Yet the optimization algorithms behind these runs have not kept pace. Most large scale training still relies on synchronous methods, where workers must wait for the slowest device, wasting compute and amplifying the effects of hardware and network variability. Removing synchronization seems like a simple fix, but asynchrony introduces staleness, meaning updates computed on outdated models. This makes analysis difficult, especially when delays arise from system level randomness rather than algorithmic choices. As a result, the time complexity of asynchronous methods remains poorly understood. This dissertation develops a rigorous framework for asynchronous first order stochastic optimization, focusing on the core challenge of heterogeneous worker speeds. Within this framework, we show that with proper design, asynchronous SGD can achieve optimal time complexity, matching guarantees previously known only for synchronous methods. Our first contribution, Ringmaster ASGD, attains optimal time complexity in the homogeneous data setting by selectively discarding stale updates. The second, Ringleader ASGD, extends optimality to heterogeneous data, common in federated learning, using a structured gradient table mechanism. Finally, ATA improves resource efficiency by learning worker compute time distributions and allocating tasks adaptively, achieving near optimal wall clock time with less computation. Together, these results establish asynchronous optimization as a theoretically sound and practically efficient foundation for distributed learning, showing that coordination without synchronization can be both feasible and optimal.

Summary and Contributions: Update following the the author response: I thank the authors for the clarifications. This paper explores the connection between learning with approximate privacy and online learning for multi-class and regression problems. It follows prior work that showed an equivalence between approximate privacy and online learning for binary classification. This paper has two main results: (1) Multi-class: Consider a hypothesis class H of functions from X to Y where Y is finite. Then, H is PAC learnable with respect to the 0-1 loss in the online setting with if and only if it is learnable with approximate privacy in the standard stochastic batch setting.

In our work, we provided a complete answer to this fundamental question by improving uniform convergence rate from O( S \sqrt{log(m)/m}) to O(\sqrt{log S }/\sqrt{m}). While it appears that we have only shaved the extra log(m) term to obtain an optimal rate, the dependence on S is improved from linear to logarithmic. As discussed in detail in Remark 1(ii, iii), the logarithmic dependence on S is optimal and it ensures that the uniform convergence guarantee can be obtained not just over fixed compact set S but over entire R d by growing S to infinity at an exponential rate, i.e., S_m o(e m) rather than at a sublinear rate, i.e., S_m o(\sqrt{m/log(m)}). In other words, for the same approximation error, the kernel can be approximated uniformly over a significantly larger S than it was considered in the literature.

Proposals of artificial intelligence (AI) solutions based on increasingly complex and accurate predictive models are becoming ubiquitous across many disciplines. As the complexity of these models grows, transparency and users' understanding often diminish. This suggests that accurate prediction alone is insufficient for making an AI-based solution truly useful. In the development of healthcare systems, this introduces new issues related to accountability and safety. Understanding how and why an AI system makes a recommendation may require complex explanations of its inner workings and reasoning processes. Although research on explainable AI (XAI) has significantly increased in recent years and there is high demand for XAI in medicine, defining what constitutes a good explanation remains ad hoc, and providing adequate explanations continues to be challenging. To fully realize the potential of AI, it is critical to address two fundamental questions about explanations for safety-critical AI applications, such as health-AI: (1) What is an explanation in health-AI? and (2) What are the attributes of a good explanation in health-AI? In this study, we examined published literature and gathered expert opinions through a two-round Delphi study. The research outputs include (1) a definition of what constitutes an explanation in health-AI and (2) a comprehensive list of attributes that characterize a good explanation in health-AI.

Note: 4.2/5 (14 notes) 6,615 students Are you Preparing for Interview in Artificial Intelligence? Don't be stressed, take our Artificial Intelligence based quiz and prepare yourself for your Interview. With this Artificial Intelligence based Quiz, we are going to build your confidence by providing tips and trick to solve Artificial Intelligence based questions. In Artificial Intelligence based Multiple Choice Questions Quiz, there will be a series of practice tests where you can test your Basic Knowledge in Artificial Intelligence. Who should Practice these Artificial Intelligence based Questions?

Three new research projects at the University of Warwick that will investigate new avenues for a sustainable future as well as a fundamental question of our universe's past have been awarded a total of more than €5 million in European Research Council Starting Grants. Following the first call for proposals under the EU's new R&I programme, Horizon Europe, €619 million will be invested in excellent projects dreamed up by 397 scientists and scholars. Grants worth on average €1.5 million will help ambitious younger researchers launch their own projects, form their teams and pursue their best ideas. The selected proposals cover all disciplines of research, from the medical applications of artificial intelligence, to the science of controlling matter by using light, to designing a legal regime for fair influencer marketing. The SHINE project (Shining Light on Metal Halide Perovskite Stability with Nanoscale Optical Microscopy and Ultrafast Spectroscopy), led by Dr Rebecca Milot of the University of Warwick's Department of Physics, has received €2,473,363 and will investigate one of the most promising new materials for solar energy conversion, metal halide perovskites.

Are you Preparing for Interview in Robotic Process Automation? Don't be stressed, take our Robotic Process Automation based quiz and prepare yourself for your Interview. With this Robotic Process Automation based Quiz, we are going to build your confidence by providing tips and trick to solve Robotic Process Automation based questions. In Robotic Process Automation based Multiple Choice Questions Quiz, there will be a series of practice tests where you can test your Basic Knowledge in Robotic Process Automation. Who should Practice these Robotic Process Automation based Questions?

Are you Preparing for Interview in Robotics? Don't be stressed, take our Robotics based quiz and prepare yourself for your Interview. With this Robotics based Quiz, we are going to build your confidence by providing tips and trick to solve Robotics based questions. In Robotics based Multiple Choice Questions Quiz, there will be a series of practice tests where you can test your Basic Knowledge in Robotics. Use it for training and keep yourself updated.

Learn more about the book here and watch Shane's presentation on the book here. Learn more about the book here and watch Russel's presentation on the book here. Moving away from the more conceptual books listed, The Hundred-Page Machine Learning Book provides a concise and practical look at the most fundamental questions in ML. In this book, Interpretable Machine Learning Researcher and Ph.D, Christoph Molnar focuses on ML's biggest issue with adoption: that these systems seldom explain their inner-workings meaning a great deal of a machine's processes are hidden within a black box. A free online version of the book can be found here, and author's presentation of the main idea from this book can be found here.