Viraj Kulkarni

Over the last 75 years, many scientists, engineers and entrepreneurs have told us again and again that intelligent computers that can really think are just around the corner. Looking at the enormous hype the field enjoys today, one would be hard-pressed to not believe it. This is inevitable and only a matter of time. But, is it?

I review Erik Larson's new book 'The Myth of Artificial Intelligence' for The Wire Science.

The use of machine learning to develop intelligent software tools for interpretation of radiology images has gained widespread attention in recent years. The development, deployment, and eventual adoption of these models in clinical practice, however, remains fraught with challenges. In this paper, we propose a list of key considerations that machine learning researchers must recognize and address to make their models accurate, robust, and usable in practice.

We are in the middle of what the journal Nature has called the “quantum gold rush”. Governments around the world are ramping up their investments in quantum computing. Venture capitalists are pouring billions of dollars into startups sprouting out of university departments. Established technology companies like IBM, Google, Microsoft, Intel, Amazon and Honeywell have recruited highly qualified teams to build quantum computers.

Discrimination is as old as humankind; religious preaching, moral education, processes or legislation may mitigate its consequences but can’t eliminate it altogether. But today, as we increasingly cede decision-making to AI algorithms, we have a unique opportunity. For the first time in history, we have a real shot at building a fair society that is free of human prejudices by building machines that are fair by design.

Time is not real. There is nothing special about the present moment; in fact, a universal present moment does not even exist. The past and the future are equal in all respects. Our notion that time flows irreversibly from the past into the future is an illusion born out of our ignorance about the world. It exists only in our subjective perceptions and not as part of objective reality. Let me convince you of this using simple mathematics, 15-mins of patience, and an open mind.

Einstein, Bohr, Schrödinger, Heisenberg and others were deeply troubled by the implications of the quantum theory they had developed. They were faced with a personal dilemma: to believe a preposterous theory that worked or discard it for an intuitive theory that didn’t work. Discovering echoes of their theory in the ancient Upanishads gave them comfort, courage, and spiritual guidance.

Present AI systems suffer some obvious limitations. They are brittle, incapable of solving problems that deviate even slightly from what they were designed for, and they are data-hungry. Critics use variations of these limitations to conclude that there exists a fundamental difference between human intelligence and artificial intelligence. This, however, may be a premature conclusion. If we look closer, it turns out that humans also suffer from these same limitations.

In the popular German legend, the protagonist Faust strikes a deal with the Devil and trades his soul for unlimited knowledge and pleasure. In today’s age of the Internet and artificial intelligence, we have struck an equally dangerous bargain: we have allowed algorithms to invade our privacy, excavate our minds and manipulate our deepest thoughts, even those we may have kept hidden from ourselves, in exchange for free entertainment and social media. We have mindlessly relinquished our freedom of thought to voyeurist profit-seeking corporations and power-hungry governments.

AI is not only making TB diagnosis accessible to people who were previously deprived of it, but by reducing radiologist effort required in reading X-rays, it is lowering healthcare costs and improving reporting times.

Science is about finding explanations. With the abundance of data, we have stopped asking for explanations and are satisfied with mere correlations. We are happy to know something will happen without knowing why it will happen. This is damaging the spirit of scientific enquiry in a fundamental way from which we may never recover.

GPT-3 is a significant achievement that pushes the boundaries of AI research in natural-language processing. OpenAI has demonstrated that, when it comes to AI, bigger is in fact better. GPT-3 uses the same architectural framework as GPT-2 but performs markedly better owing only to its size. This leads us to an important question: can the limitations of GPT-3 be overcome simply by throwing more data and computational horsepower at it?

Quantum computers can benefit machine learning research and application across all science and engineering domains. In this paper, we provide a background and summarize key results of quantum computing before exploring its application to supervised machine learning problems. By eschewing results from physics that have little bearing on quantum computation, we hope to make this introduction accessible to data scientists, machine learning practitioners, and researchers from across disciplines.

Quantitative trading is a melting cauldron of computer programming, machine learning, finance, and statistics. To be a quant, you need to be a jack of all these trades and a master of at least some of them. Regardless of the discipline you come from and the approaches you use, there are three things you must absolutely take care of which often leave even experienced traders confounded.

Tired of fulfilling endless demands of people around her and frustrated by society's coercive attempts to mould her according to its expectations, young Asmi abandons her mundane life and embarks on a remarkable journey to seek freedom and happiness. Her life changes when an unusual, contemporary monk introduces her to the timeless wisdom of the Upanishads. As she delves into the practice of Vedanta and meditation, she begins to unravel the mysteries of the meaning of 'I', the nature of reality, and the causes of human bondage.

Federated learning enables machine learning models to learn from private decentralized data without compromising privacy. The standard formulation of federated learning produces one shared model for all clients. Due to statistical heterogeneity in data, the local models trained solely on a client's private data may perform better than the global shared model. This paper highlights the need for subsequent model personalization and surveys recent research on this topic.