Chip design is a long slog of trial and error, taking years to bring a design to market. Motivo, a five year old startup from a chip industry veteran is creating software to speed up chip design from years to months using AI. Today the company announced a $12 million Series A. Intel Capital led the round along with new investors Storm Ventures and Seraph Group, as well as participation from Inventus Capital. The company reports it has now raised a total of $20 million with its previous seed funding. Motivo co-founder and CEO Bharath Rangarajan has worked in the chip industry for 30 years, and he saw a few fundamental trends and issues.

Does Google Assistant always say your name wrong, or maybe the names of people you know? You can soon teach the digital assistant how to pronounce them correctly. Google announced an update rolling out soon to Assistant, available on smartphones and Google Home speakers, that will allow users to teach it how to properly pronounce your name or those in your contacts. Google said the feature will initially be available in English but will roll out to offer more languages soon. "Names matter, and it's frustrating when you're trying to send a text or make a call and Google Assistant mispronounces or simply doesn't recognize a contact," said Yury Pinsky, director of product management at Google, in a blog post published Wednesday.

Can sensors see behind the corners of obstacles in real time? As it turns out, yes. A study by researchers at Stanford, Rice University, Princeton, and Southern Methodist University published in the journal Optica proposes a system that's capable of producing around-the-bend images at high resolutions and speeds. It's able to distinguish the submillimeter details of hidden objects from 1 meter away, and according to coauthor Felix Heide, it could be used to make out things like the license plates of hidden moving vehicles and personnel badges worn by walking individuals. "Non-line-of-sight imaging has important applications in medical imaging, navigation, robotics, and defense," said Heide.

Speaking to Rangarajan Sundaram, who began his tenure as dean of NYU Stern School of Business in January this year, you don't get any sense of the weight that sits on his shoulders. After all, "it's a wonderful time to be leading the most exciting business school in the world," he says. Leading such a revered institution requires constant innovation, as the world around us is challenged to adapt to the rapid way machine learning, big data, and the future of tech are changing the way we learn. The transformation of the NYU Stern MBA reflects this. For Rangarajan, his school, and business schools like it, have a responsibility to ensure the next generation of business leaders aren't just part of the change, but lead it.

This paper describes a simple geometrical Concave-Convex procedure (CCCP) for constructing discrete time dynamical systems which can be guaranteed to decrease almost any global optimization/energy function (see technical conditions in section (2)). We prove that there is a relationship between CCCP and optimization techniques based on introducing auxiliary variables using Legendre transforms. We distinguish between Legendre min-max and Legendre minimization. In the former, see [6], the introduction of auxiliary variables converts the problem to a min-max problem where the goal is to find a saddle point. By contrast, in Legendre minimization, see [8], the problem remains a minimization one (and so it becomes easier to analyze convergence).

This paper describes a simple geometrical Concave-Convex procedure (CCCP) for constructing discrete time dynamical systems which can be guaranteed to decrease almost any global optimization/energy function (see technical conditions in section (2)). We prove that there is a relationship between CCCP and optimization techniques based on introducing auxiliary variables using Legendre transforms. We distinguish between Legendre min-max and Legendre minimization. In the former, see [6], the introduction of auxiliary variables converts the problem to a min-max problem where the goal is to find a saddle point. By contrast, in Legendre minimization, see [8], the problem remains a minimization one (and so it becomes easier to analyze convergence).

We introduce the Concave-Convex procedure (CCCP) which constructs discretetime iterative dynamical systems which are guaranteed to monotonically decrease global optimization/energy functions. It can be applied to (almost) any optimization problem and many existing algorithms can be interpreted in terms of CCCP. In particular, we prove relationships to some applications of Legendre transform techniques. We then illustrate CCCP by applications to Potts models, linear assignment, EM algorithms, and Generalized Iterative Scaling (GIS). CCCP can be used both as a new way to understand existing optimization algorithms and as a procedure for generating new algorithms. 1 Introduction There is a lot of interest in designing discrete time dynamical systems for inference and learning (see, for example, [10], [3], [7], [13]).

The softassign quadratic assignment algorithm has recently emerged as an effective strategy for a variety of optimization problems in pattern recognition and combinatorial optimization. While the effectiveness of the algorithm was demonstrated in thousands of simulations, there was no known proof of convergence. Here, we provide a proof of convergence for the most general form of the algorithm.

The softassign quadratic assignment algorithm has recently emerged as an effective strategy for a variety of optimization problems in pattern recognition and combinatorial optimization. While the effectiveness of the algorithm was demonstrated in thousands of simulations, there was no known proof of convergence. Here, we provide a proof of convergence for the most general form of the algorithm.

The softassign quadratic assignment algorithm has recently emerged as an effective strategy for a variety of optimization problems inpattern recognition and combinatorial optimization. While the effectiveness of the algorithm was demonstrated in thousands of simulations, there was no known proof of convergence. Here, we provide a proof of convergence for the most general form of the algorithm.