Can the success of reinforcement learning methods for combinatorial optimization problems be extended to multi-robot scheduling problems in stochastic contexts? Three issues are particularly important in this context: quality of the resulting decisions, scalability, and transferability. To achieve these ends we generalize the concept of clique potential to stochastic clique potential. We extend a mean field inference fixed point iteration with this new concept and use it to modify thestructure2vec method. We next propose a new reinforcement learning framework combining a graph representation of the problem and a consensus auction inspired by heuristics in the problem domain. This representation enables transferability in terms of the number of robots. Sequential encoding of information through multiple layers of our extended structure2vec results in 96% optimal performance of the learned heuristics. While training tractability is inherited from single robot methods in the literature, use of a multi-robot consensus auction-based relaxation of the maximum operation in the Bellman optimality equation allows for scalable selection of actions in the fitted Q-iteration. We apply our framework to multi-robot reward collection (MRRC) problems in stochastic environments with linear or non-linear rewards. In stochastic environments with non-linear rewards, the new method achieves 20% superior performance relative to the popular sequential greedy assignment (SGA) algorithm. Linear scalability in terms of training is achieved and demonstrated. Transferability is demonstrated by the use of a heuristic trained with three robots that continues to achieve 95% optimal performance when applied to problems with various numbers of robots. We further mention the results obtained when extending the approach to identical parallel machine scheduling(IPMS) problems.

The GLUE benchmark (Wang et al., 2019b) is a suite of language understanding tasks which has seen dramatic progress in the past year, with average performance moving from 70.0 at launch to 83.9, state of the art at the time of writing (May 24, 2019). Here, we measure human performance on the benchmark, in order to learn whether significant headroom remains for further progress. We provide a conservative estimate of human performance on the benchmark through crowdsourcing: Our annotators are non-experts who must learn each task from a brief set of instructions and 20 examples. In spite of limited training, these annotators robustly outperform the state of the art on six of the nine GLUE tasks and achieve an average score of 87.1. Given the fast pace of progress however, the headroom we observe is quite limited. To reproduce the data-poor setting that our annotators must learn in, we also train the BERT model (Devlin et al., 2019) in limited-data regimes, and conclude that low-resource sentence classification remains a challenge for modern neural network approaches to text understanding.

Human thinking requires the brain to understand the meaning of language and to properly organize the thoughts flow using the language. However, current natural language processing models are primarily limited to the word probability estimation. Here, we proposed a Language Guided Imagination (LGI) network to incrementally learn the meaning and usage of diverse words and syntaxes, aiming to form a humanlike machine thinking process. LGI contains three subsystems: (1) vision system that contains an encoder to disentangle the input or imagined scenarios into abstract population representations, and an imagination decoder to reconstruct imagined scenarios from higher level representations; (2) Language system, which consists of a binarizer to transfer symbol texts into binary vectors, an IPS (mimicking the human IntraParietal Sulcus, implemented by an LSTM) to extract the quantity information from the input texts, and a textizer to convert binary vectors into text symbols; (3) a PFC (mimicking the human PreFrontal Cortex, implemented by an LSTM) that combines inputs in the forms of both language and vision, and predict text symbols and manipulated images accordingly. In this work, the proposed LGI network illustrates the ability to incrementally learn eight different syntaxes and form a machine thinking loop that enables interactions between language and vision system, which hasn't been demonstrated before. The paper presents a new architecture that allows the machine to learn, understand and use language in a humanlike way, which might ultimately enable a machine to construct fictitious'mental' scenario and possess intelligence.

A new machine learning model allows for physicians to determine whether atypical ductal hyperplasia (ADH) could upgrade to cancer, according to new research published in JCO Clinical Cancer Informatics. The model can identify 98 percent of all malignant cases prior to surgery, while sparing 16 percent of women from unnecessary surgeries on benign lesions. ADH is a breast lesion that increases the risk of breast cancer by four- to five-fold. Typically, ADH is found with mammography and its presence confirmed using biopsy. Previous research published in Current Problems in Diagnostic Radiology found that 95 percent of breast imagers recommend surgical removal for all ADH cases discovered during biopsy to establish if the lesion is cancerous.

The idea is to "anticipate, absorb and recover from events that cause grid outages, such as extreme weather or a cyberattack," said Ashley Pilipiszyn, GRIP project lead and a Ph.D. student at Stanford University. The project is co-led by the SLAC National Accelerator Laboratory, which is operated by Stanford University, and the Lawrence Berkeley National Laboratory, managed by the University of California. Like many such initiatives focused on artificial intelligence and climate change, the public and private sectors are involved in supplying research and funds. In the case of a failure caused by a winter storm, for example, Ms. Pilipiszyn said that a smart grid could prioritize different electrical loads into islands and isolate faults, ensuring, say, that a nursing home or hospital receives top priority. GRIP is a three-year project, and field demonstrations are expecting to be up and running by the end of 2020, Ms. Pilipiszyn said.

The fear of the future can be a powerful deterrent to change, as shown by a recent McKinsey Global Survey about the spread of automation in the workplace. Almost half of the respondent executives who rated their current automation programs a success nevertheless identified two factors--managing employee resistance to change and attracting talent--as their biggest challenges to adopting automation over the next three years (Exhibit 1). It's a truism that change brings challenges. But it's equally true that change brings opportunities. Automation has huge potential to change the nature of work, most obviously by freeing up workers from repetitive and tedious tasks.

Artificial intelligence (AI) has the potential to help tackle some of the world's most challenging social problems. To analyze potential applications for social good, we compiled a library of about 160 AI social-impact use cases. They suggest that existing capabilities could contribute to tackling cases across all 17 of the UN's sustainable-development goals, potentially helping hundreds of millions of people in both advanced and emerging countries. Real-life examples of AI are already being applied in about one-third of these use cases, albeit in relatively small tests. They range from diagnosing cancer to helping blind people navigate their surroundings, identifying victims of online sexual exploitation, and aiding disaster-relief efforts (such as the flooding that followed Hurricane Harvey in 2017). AI is only part of a much broader tool kit of measures that can be used to tackle societal issues, however. For now, issues such as data accessibility and shortages of AI talent constrain its application for social good. This article is a condensed version of our discussion paper, Notes from the AI frontier: Applying AI for social good (PDF–3MB).

Today, the contact center is at an inflexion point. Artificial intelligence is here and with it comes the ability to recognize customer situations and solve their problems faster. With the rise of chatbots, this marks the beginning of the shift to digital labor, meaning the automation of tasks that are performed by computer applications that were previously performed by humans. Digital labor can be used in contact centers to solve problems that humans are having with a particular product or service. Aragon predicts that by 2021, digital labor will become a key feature of intelligent contact center offerings.

How did you learn to talk? Probably something like this: Your infant brain, a hotbed of neurological activity, picked up on your parents' speech tones and facial expressions. You started to mimic their sounds, interpret their emotions and identify relatives from strangers. And one day, about a year into life, you pointed and started saying a few meaningful words with slobbery glee. But many children, particularly those diagnosed with autism spectrum disorder, acquire language in different ways.

In the past five years, machine learning has come a long way. You might have noticed that Siri, Alexa, and Google Assistant are way better than they used to be, or that automatic translation on websites, while still fairly spotty, is hugely improved from where it was a few years ago. But many still don't quite grasp how far we've come, and how fast. Recently, two images made the rounds that underscore the huge advances machine learning has made -- and show why we're in for a new age of mischief and online fakery. The first was put together by Ian Goodfellow, the director of machine learning at Apple's Special Projects Group and a leader in the field. He looked over machine-learning papers published on the online open-access repository arXiv over the past five years, and found examples of machine learning-generated faces from each year.