LONDON--(BUSINESS WIRE)--Deep Knowledge Analytics, Big Innovation Centre and Innovation Eye launch'Artificial Intelligence in the UK: Industry landscape overview in 2021', the most comprehensive industry mapping made to date, profiling, categorising and analysing over 3,600 private and public sector entities across 20 sectors and 50 locations in the UK. The open-access report and IT platform, covers the latest developments in technology and innovation, best ranking companies and investors hubs, AI and COVID-19, policy and ethics, AI challenges and opportunities for the UK. Additionally, it profiles the top 100 UK AI experts and hubs, including think tanks, tech-hubs and doctoral training centres. The UK remains first in Europe and ranks third globally, behind the USA and China in developing AI technologies with a reported £9 billion investment growth for the industry between 2019 and 2021. The UK government continues to foster growth and industry development through initiatives in 5 key areas - human capital, lab to market developments, networking, regulation and infrastructure.

Common sense isn't common, especially when it comes to artificial intelligence. Computers struggle to make fine distinctions that people take for granted. This is why websites require you authenticate your humanity before logging in or making a purchase: Most bots can't tell the difference between a crosswalk and a zebra. At the USC AI Futures Symposium on AI with Common Sense earlier this month, more than 20 USC researchers reported on the technical reasons why that's the case, and different avenues of research to address this. Advances in common sense AI will improve human-facing services, from enhanced social services to better serve society to personal assistants that better predict our context and needs.

The United States on Thursday unleashed a volley of actions to censure China's treatment of the Uyghur minority, with lawmakers voting to curb trade and new sanctions slapped on the world's top consumer drone maker. The United States has been ramping up pressure on China amid a crop of disputes, with President Joe Biden's administration a day earlier targeting producers of painkillers that have contributed to America's addiction crisis. The U.S. Senate unanimously voted to make the United States the first country to ban virtually all imports from China's northwestern Xinjiang region over concerns of the prevalence of forced labor. "We know it's happening at an alarming, horrific rate with the genocide that we now witness being carried out," said Senator Marco Rubio, a driver behind the act, which already passed the House of Representatives and which the White House says Biden will sign. After prolonged negotiations to secure its passage, Rubio lifted objections and the Senate confirmed veteran diplomat Nicholas Burns as ambassador to China.

Google parent company Alphabet is folding one of its subsidiaries back into Google as the startup's founder steps down to confront a neurological disease. Sidewalk Labs CEO Dan Doctoroff said Thursday he's leaving the company focused on environmentally sustainable urban planning technology because he "very likely" has ALS, also known as Lou Gehring's disease. Sidewalk Labs was one of a hodgepodge of projects to spin off from Google when the tech giant put itself under a new holding company, Alphabet, in 2015. The idea was to separate Google's riskier explorations of futuristic technology from its highly profitable, advertising-fueled core business. Doctoroff said in a Medium post about his departure Thursday that four products developed by Sidewalk Labs now fit squarely into Google's commitments to be more environmentally sustainable and go carbon-free by 2030.

NSF's financial staff developed an artificial intelligence application called the Financial Obligation Root Cause Explorer, or FORCE.

Utilidata, an industry leading grid-edge software company, announced today that it is developing a software-defined smart grid chip in collaboration with NVIDIA. The chip will be powered by NVIDIA's AI platform and embedded in smart meters to enhance grid resiliency, integrate distributed energy resources (DERs) -- including solar, storage, and electric vehicles (EVs) -- and accelerate the transition to a decarbonized grid. The U.S. Department of Energy's (DOE's) National Renewable Energy Laboratory (NREL) will be among the first to test the software-defined smart grid chip as a way to scale and commercialize the lab's Real-Time Optimal Power Flow (RT-OPF) technology, with support from the Solar Energy Technologies Office Technology Commercialization Fund. Originally developed with funding from DOE's Advanced Research Projects – Energy (ARPA-E) program, RT-OPF enables highly localized load control to seamlessly integrate an increasing number of DERs while ensuring stable and efficient grid operations. "To date, the scalability and commercial potential of technologies like RT-OPF have been limited by single-use hardware solutions," said Santosh Veda, Group Manager for Grid Automation and Controls at NREL. "By developing a smart grid chip that can be embedded in one of the most ubiquitous utility assets – the smart meter – this approach will potential enable wider adoption and commercialization of the technology and redefine the role of edge computing for DER integration and resiliency. Enhanced situational awareness and visibility from this approach will greatly benefit both the end customers and the utility."

We fine-tune GPT-3 to answer long-form questions using a text-based web-browsing environment, which allows the model to search and navigate the web. By setting up the task so that it can be performed by humans, we are able to train models on the task using imitation learning, and then optimize answer quality with human feedback. To make human evaluation of factual accuracy easier, models must collect references while browsing in support of their answers. We train and evaluate our models on ELI5, a dataset of questions asked by Reddit users. Our best model is obtained by fine-tuning GPT-3 using behavior cloning, and then performing rejection sampling against a reward model trained to predict human preferences. This model's answers are preferred by humans 56% of the time to those of our human demonstrators, and 69% of the time to the highest-voted answer from Reddit.

Reinforcement learning (RL) is a central problem in artificial intelligence. This problem consists of defining artificial agents that can learn optimal behaviour by interacting with an environment -- where the optimal behaviour is defined with respect to a reward signal that the agent seeks to maximize. Reward machines (RMs) provide a structured, automata-based representation of a reward function that enables an RL agent to decompose an RL problem into structured subproblems that can be efficiently learned via off-policy learning. Here we show that RMs can be learned from experience, instead of being specified by the user, and that the resulting problem decomposition can be used to effectively solve partially observable RL problems. We pose the task of learning RMs as a discrete optimization problem where the objective is to find an RM that decomposes the problem into a set of subproblems such that the combination of their optimal memoryless policies is an optimal policy for the original problem. We show the effectiveness of this approach on three partially observable domains, where it significantly outperforms A3C, PPO, and ACER, and discuss its advantages, limitations, and broader potential.

When legal regulations get ahead of technological developments, the progress of human society may be constrained. When technological developments run ahead of legal regulations, the unregulated new technologies may harm instead of benefit human society, defying technological development's fundamental purpose. This is exactly what has happened in our world in the past decade, as technological developments far outpaced legal regulations. Worse, traditional legal frameworks focus on the relation between people, whereas we must develop a legal framework to regulate relations between people and intelligent machines in the current era. Integrating AI technologies into human society imposes unique legal challenges without any precedence.

Published in June 2009 at a computer vision conference in Florida, ImageNet's open dataset quickly became the basis of an annual challenge to see which algorithm would have the lowest error rate in identifying images.2 In the inaugural competition, held in 2010, every team had an error rate of at least 25%. However, by combining the techniques of deep learning with the massive set of training data available with ImageNet, researchers sent error rates tumbling. By 2017, the last year of the competition, the error rate was less than 3%.3 ImageNet provided a big boost to AI--the dataset is credited with the resurgence of deep learning.4