If you are looking for an answer to the question What is Artificial Intelligence? and you only have a minute, then here's the definition the Association for the Advancement of Artificial Intelligence offers on its home page: "the scientific understanding of the mechanisms underlying thought and intelligent behavior and their embodiment in machines."
However, if you are fortunate enough to have more than a minute, then please get ready to embark upon an exciting journey exploring AI (but beware, it could last a lifetime) …
The Stanford research who made headlines in 2017 for designing an AI that uses'facial landmarks' to determine a person's sexual preference is back with what may be another controversial system. Dr. Michal Kosinski claims to have a facial recognition algorithm capable of identifying if a person is a liberal or conservative based on a single photo – and with over 70 percent accuracy. The technology, which builds on the 2017 AI, was trained with more than a million images from dating websites and Facebook and programmed to focus in on expressions and posture. Although Kosinski and his team were unable to pin down exact characteristics the algorithm associated with a political preference, but they did find some trends like head orientation and emotional expression in pictures. Some examples include people who looked directly at the camera were labeled as liberal and those showing disgust were judged as more conservative.
A study was conducted on the business adoption of Artificial Intelligence (AI) in the 1980s. Published in the MIS Quarterly, the study found that enterprises were rushing to invest in AI, and the projected market value was $4 billion. However, the results were shocking. The study found that over a five year period, just 33% of AI solutions delivered business value, while the rest were abandoned. Many popular applications of AI were proven to be pure hype and several companies became disillusioned with AI.
The coronavirus test being provided daily to tens of thousands of residents in Los Angeles and other parts of California may be producing inaccurate results, according to a warning from federal officials that could raise questions about the accuracy of infection data shaping the pandemic response. The guidance from the Food and Drug Administration warns healthcare providers and patients that the test made by Curative, a year-old Silicon Valley start-up that supplies the oral-swab tests at L.A.'s 10 drive-through testing sites, carries a "risk of false results, particularly false negative results." To reduce the risk of false negatives, the Curative test should be used only on "symptomatic individuals within 14 days of COVID-19 symptom onset," and the swab should be observed and directed by a healthcare worker, the FDA said. The guidance, issued Monday, repeats the instructions that the FDA issued when the test was first granted an emergency-use authorization. The FDA warning appears to sharply contradict Los Angeles Mayor Eric Garcetti, who in April made coronavirus testing available to anyone, regardless of symptoms.
The US Department of Homeland Security (DHS) is piloting facial recognition technologies that can see through face masks with a "promising" level of accuracy, meaning that travelers could end up breezing through airports without the need to uncover their mouths and noses at border checks. The trials were organized as part of a yearly biometric technology rally, organized by the Science and Technology Directorate, which is the research and development unit within the DHS. Every year since 2018, the rally brings together experts, technology vendors and volunteers to test top-notch biometric systems, and make sure that they are up to the challenges posed by the use of facial recognition technology in a variety of scenarios. This year, in response to the new imperatives brought by the Covid-19 pandemic, the rally has focused on evaluating the ability of AI systems to reliably collect and match images of individuals wearing an array of different face masks, with a view of eventually deploying the technology in international airports around the country. During a ten-day event, 60 facial recognition configurations were tested with the help of almost 600 volunteers from 60 different countries.
Researchers at the University of California San Diego have devised an artificial intelligence (AI) tool to predict the level of loneliness in adults, with 94% accuracy. The tool used Natural Language Processing (NLP) developed by IBM to process large amounts of unstructured natural speech and text data. It analysed factors like cognition, mobility, sleep and physical activity to understand the process of aging. This tool is an example of how AI can be used in devices to detect mental health conditions. Market research firm Gartner predicts, by 2022, your personal device will know more about your emotional state than your own family members.
A decade ago, Chatroulette was an internet supernova, exploding in popularity before collapsing beneath a torrent of male nudity that repelled users. Now, the app, which randomly pairs strangers for video chats, is getting a second chance, thanks in part to a pandemic that has restricted in-person social contact, but also thanks to advances in artificial intelligence that help filter the most objectionable images. User traffic has nearly tripled since the start of the year, to 4 million monthly unique visitors, the most since early 2016, according to Google Analytics. Founder and chairman Andrey Ternovskiy says the platform offers a refreshing antidote of diversity and serendipity to familiar social echo chambers. On Chatroulette, strangers meet anonymously and don't have to give away their data or wade through ads.
It has been only two weeks into the last month of the year and arxiv.org, the popular repository for ML research papers has already witnessed close to 600 uploads. This should give one the idea of the pace at which machine learning research is proceeding; however, keeping track of all these research work is almost impossible. Every year, the research that gets maximum noise is usually from companies like Google and Facebook; from top universities like MIT; from research labs and most importantly from the conferences like NeurIPS or ACL. In this article, we have compiled a list of interesting machine learning research work that has made some noise this year. This is the seminal paper that introduced the most popular ML model of the year -- GPT-3.
MLOps refers to machine learning operations. It is a practice that aims to make machine learning in production efficient and seamless. While the term MLOps is relatively nascent, it draws comparisons to DevOps in that it's not a single piece of technology but rather a shared understanding of how to do things the right way. The shared principles MLOps introduces encourage data scientists to think of machine learning not as individual scientific experiments but as a continuous process to develop, launch, and maintain machine learning capabilities for real-world use. Machine learning should be collaborative, reproducible, continuous, and tested. The practical implementation of MLOps involves both adopting certain best practices and setting up an infrastructure that supports these best practices.
The multiplayer online battle arena (MOBA) games are becoming increasingly popular in recent years. Consequently, many efforts have been devoted to providing pre-game or in-game predictions for MOBA games. However, these works are limited in the following two aspects: 1) the lack of sufficient in-game features; 2) the absence of interpretability in the prediction results. These two limitations greatly restrict their practical performances and industrial applications. In this work, we collect and release a large-scale dataset containing rich in-game features for the popular MOBA game Honor of Kings. We then propose to predict four types of important events in an interpretable way by attributing the predictions to the input features using two gradient-based attribution methods: Integrated Gradients and SmoothGrad. To evaluate the explanatory power of different models and attribution methods, a fidelity-based evaluation metric is further proposed. Finally, we evaluate the accuracy and Fidelity of several competitive methods on the collected dataset to assess how well do machines predict the events in MOBA games.
The attention mechanism is becoming increasingly popular in Natural Language Processing (NLP) applications, showing superior performance than convolutional and recurrent architectures. However, general-purpose platforms such as CPUs and GPUs are inefficient when performing attention inference due to complicated data movement and low arithmetic intensity. Moreover, existing NN accelerators mainly focus on optimizing convolutional or recurrent models, and cannot efficiently support attention. In this paper, we present SpAtten, an efficient algorithm-architecture co-design that leverages token sparsity, head sparsity, and quantization opportunities to reduce the attention computation and memory access. Inspired by the high redundancy of human languages, we propose the novel cascade token pruning to prune away unimportant tokens in the sentence. We also propose cascade head pruning to remove unessential heads. Cascade pruning is fundamentally different from weight pruning since there is no trainable weight in the attention mechanism, and the pruned tokens and heads are selected on the fly. To efficiently support them on hardware, we design a novel top-k engine to rank token and head importance scores with high throughput. Furthermore, we propose progressive quantization that first fetches MSBs only and performs the computation; if the confidence is low, it fetches LSBs and recomputes the attention outputs, trading computation for memory reduction. Extensive experiments on 30 benchmarks show that, on average, SpAtten reduces DRAM access by 10.0x with no accuracy loss, and achieves 1.6x, 3.0x, 162x, 347x speedup, and 1,4x, 3.2x, 1193x, 4059x energy savings over A3 accelerator, MNNFast accelerator, TITAN Xp GPU, Xeon CPU, respectively.