This post was co-written with Baptiste Rocca. In the last few years, deep learning based generative models have gained more and more interest due to (and implying) some amazing improvements in the field. Relying on huge amount of data, well-designed networks architectures and smart training techniques, deep generative models have shown an incredible ability to produce highly realistic pieces of content of various kind, such as images, texts and sounds. Among these deep generative models, two major families stand out and deserve a special attention: Generative Adversarial Networks (GANs) and Variational Autoencoders (VAEs). In a previous post, published in January of this year, we discussed in depth Generative Adversarial Networks (GANs) and showed, in particular, how adversarial training can oppose two networks, a generator and a discriminator, to push both of them to improve iteration after iteration.

Inspur, a leading data center and AI full-stack solutions provider, today announced two AI-technology driven open computing systems. The X-MAN 4.0, developed with Baidu, is the world's first OAI (Open Accelerator Infrastructure) compliant and liquid cooling rack-scale AI computing product optimized specifically for deep neural network applications. The Inspur OAI UBB system, meanwhile, is a 21-inch Full-Rack OAM solution delivering efficiency, flexibility and management. Workloads in data centers are growing more diverse and complex with artificial intelligence and other emerging technologies and applications spreading rapidly. Plus, Internet companies are struggling with AI's increasing hardware complexity--integrating an AI accelerator typically takes 6 to 12 months.

What are the core stages in the future development of artificial intelligence and how might its application evolve over the next 20 years? "Artificial intelligence (AI) is coming to get us" and it's just a small step from a self-driving car to AIs taking over our world and the "end of days" dystopias that Hollywood has popularized in a number of recent films. Or so some would have us believe. The reality is far less dramatic, and AI will need to go through a number of stages of development before it reaches the most extreme scenarios--if ever. In our recent book Beyond Genuine Stupidity--Ensuring AI Serves Humanity, we explore the possibilities and challenges presented by this game-changing technology.

In this experiment, we follow the experimental setup proposed by You et al. (2018). We optimize the penalized logP score of 800 low-scoring molecules from the ZINC data set. Our genetic algorithm is initiated with a molecule from the data set, and we run each experiment for 20 generations and a population size of 500 without the discriminator. For each run, we report the molecule m that increases the penalized logP the greatest, while possessing a similarity sim(m,m′) δ with the respective reference molecules m′. We calculate molecular similarity based on Morgan Fingerprints of radius 2. To ensure generation of molecules possessing a certain similarity, for molecule m we modify the fitness to: Here, SimilarityPenalty(m) is 0 if sim(m,m′) δ and 106 otherwise.

You can follow any comments to this entry through the RSS 2.0 feed. You can skip to the end and leave a comment. Pinging is currently not allowed. Your email address will not be published. This site uses Akismet to reduce spam.

In the years ahead we will increasingly see machine learning models move from the digital to the physical world. Whether it's autonomous vehicles skilfully navigating the chaotic streets of Brussels, or surveillance systems accurately recognising faces in jumbled crowds. For better or worse, AI is moving into our everyday lives. A very recent example of this can already be found close to home – at Colruyt. The Belgian supermarket is experimenting with a new kind of smart register.

By employing an artificial intelligence network typically used for image recognition, researchers at University of Texas at Dallas aim to skip labor-intensive processing steps while reacting to raw nerve signal data in real time.

Deep Knowledge Ventures (DKV) is a Hong Kong based investment fund with teams in London, Geneva, and San Francisco. The fund primarily invests in DeepTech, AI, and advanced biomedicine. In 2015, DKV incorporated a subsidiary investment fund Deep Knowledge Life Sciences (DKLS), a London-based venture fund, focused on disruptive biopharmaceutical, medical device, and healthcare companies in a partnership with scientists at the universities of Oxford and Cambridge. DKLS it has been the lead investor so far in a number of promising biomedicine and longevity companies: Insilico Medicine (and the consortium of companies around Insilico Medicine, including Youth Laboratories and Longenesis) and other companies in the fields of Geroscience, NeuroTech, Preventive Medicine, AgeTech and Longevity. The Pharma Division of Deep Knowledge Analytics is the leading analytical entity of DKLS specifically focused on deep intelligence of the pharma industry and the AI for Drug Discovery sector. So far its reports have been covered by top-tier tech, business and finance media -- including Forbes, Financial Times, acknowledged by many other authoritative entities such as MIT Review -- and have been recognized by industry experts and executives as the dominant source of sophisticated industry analysis, pragmatic forecasting and tangible benchmarking on the AI in Pharma sector.

Ask anyone who knows a bit about technology, and you'll more than likely hear them say a few things about machine learning and/or deep learning. These two concepts are catapulting artificial intelligence into strata never seen before, and sometimes it seems as though they are the same thing – but look closely and it becomes clear that they are two different beasts, and both must be tamed before artificial intelligence achieves its next breakthrough. At its simplest, it is possible to describe machine learning as a process by which machines use algorithms to take in data, learn from that data, and use what they've learned to make more nuanced, informed decisions. This involves a great deal of data and usually brings machine and social networking into the picture. For example, music streaming services are capable of looking at what you listen to, finding other users who listen to the same music, and suggesting more music for you based upon the preferences of users similar to yourself.