Like a failed communist state traditional machine learning centralises training of a model on a single machine. Centralising data in a single central location is not always possible for a variety of reasons such as slow network connections, and legal constraints. These limitations have produced a series of techniques that allow the decentralised training of a model. This collection of techniques is referred to as Federated Machine Learning. Read More
Monthly Archives: October 2019
Machine Learning Advances and Edge Computing Redefining IoT
Smart, connected products are changing the face of competition. That was the thesis of a formative 2014 article in Harvard Business Review that highlighted the transformative potential of information technology integrated into an array of products.
In the past five years, however, the seemingly straightforward words of “smart” and “connected” have become more enigmatic and, arguably, more loaded terms. And the meaning of those two terms has steadily evolved and continues to change. Five to ten years ago, a “smart” product was one with embedded sensors, processors and software. These days, to qualify as “smart,” a device needs to take advantage of some form of basic machine learning at a minimum. Read More
Artificial Intelligence Is on the Case in the Legal Profession
Artificial intelligence (AI) is, in fact, becoming a mainstay component of the legal profession. In some circumstances, this analytics-crunching technology is using algorithms and machine learning to do work that was previously done by entry-level lawyers. (What does that say about entry-level lawyers?)
Apparently, AI robot lawyers are here—and they’re not going away. Read More
API Is Dead – Long Live the APIs
In this article, the author will highlight how the reign of REST APIs is declining and how the ecosystem is moving towards democracy. Read More
#microservicesDespite Their Huge Upside Potential, Why Do Most Platforms Fail?
“For anyone who follows the world of business, it is now common knowledge that the most valuable firms on the planet and the first companies to surpass the trillion-dollar mark in value (albeit temporarily) are platforms,” Michael Cusumano, Annabelle Gawer and David Yoffie write in their recently published book “The Business of Platforms: Strategy in the Age of Digital Competition, Innovation, and Power.” In a list of more than 200 startups with valuations of $1 billion or more, they estimated that platforms made up between 60% and 70%.
What do we mean by platform? The book offers a simple definition. “Platforms, in general, connect individuals and organizations for a common purpose or to share a common resource.” Read More
How to tell if your industry is ready for platform transformation
Platform strategy is sweeping through the business landscape, upending industries as companies adopt a business model based on enabling interactions between external producers and consumers. Companies like Airbnb, Uber, YouTube, Amazon, and Facebook have set the model for platform economies, with competing companies scrambling to catch up.
Not all industries have been affected equally, however. Retail executives and journalists have seen their industries transform rapidly, while airplane engineers and heart surgeons have not.
Platform transformations work best in industries with precise output, low regulation, and spare capacity. Read More
Platform strategy, explained
Platforms are environments, computing or otherwise, that connect different groups and derive benefits from others participating in the platform. The underlying concept covers companies from Google to Facebook to video game platform Steam to Taser (more on that later).
“’Platform Strategy’ is one of our few courses where participants can spend an hour debating on what they are learning about is,” MIT Sloan Professor Catherine Tucker tells students in her executive education course on platform strategy. “Don’t get hung up on definitions. Being a platform or not is more of a range than a set point.” Read More
The 7 levels of the Internet of Things
Is it possible to build a secure, collaborative platform required to support IoT’s solutions orientation? Cisco believes it is, and outlined a first step in this direction at the event with announcement of the IoT Reference Model. Read More
Solving Rubik’s Cube with a Robot Hand
We demonstrate that models trained only in simulation can be used to solve a manipulation problem of unprecedented complexity on a real robot. This is made possible by two key components: a novel algorithm, which we call automatic domain randomization (ADR) and a robot platform built for machine learning. ADR automatically generates a distribution over randomized environments of ever-increasing difficulty. Control policies and vision state estimators trained with ADR exhibit vastly improved sim2real transfer. For control policies, memory-augmented models trained on an ADR-generated distribution of environments show clear signs of emergent meta-learning at test time. The combination of ADR with our custom robot platform allows us to solve a Rubik’s cube with a humanoid robot hand, which involves both control and state estimation problems. Videos summarizing our results are available: https://openai.com/blog/solving-rubiks-cube/ Read More
Adversarial Attacks on Deep Neural Networks: an Overview
Deep Neural Networks are highly expressive machine learning networks. Researchers have found that it is far too easy to fool them with an imperceivable, but carefully constructed nudge in the input. Adversarial training looks to defend against attacks by pretending to be the attacker, generating a number of adversarial examples against your own network, and then explicitly training the model to not be fooled by them. Defensive distillation looks to train a secondary model whose surface is smoothed in the directions an attacker will typically try to exploit, making it difficult for them to discover adversarial input tweaks that lead to incorrect categorization. Read More
Rick's Cafe AI 