Our work reveals a structured shortcoming of the existing mainstream self-supervised learning methods. Whereas self-supervised learning frameworks usually take the prevailing perfect instance level invariance hypothesis for granted, we carefully investigate the pitfalls behind. Particularly, we argue that the existing augmentation pipeline for generating multiple positive views naturally introduces out-of-distribution (OOD) samples that undermine the learning of the downstream tasks. Generating diverse positive augmentations on the input does not always pay off in benefiting downstream tasks. To overcome this inherent deficiency, we introduce a lightweight latent variable model UOTA, targeting the view sampling issue for self-supervised learning. UOTA adaptively searches for the most important sampling region to produce views, and provides viable choice for outlier-robust self-supervised learning approaches. Our method directly generalizes to many mainstream self-supervised learning approaches, regardless of the loss's nature contrastive or not. We empirically show UOTA's advantage over the state-of-the-art self-supervised paradigms with evident margin, which well justifies the existence of the OOD sample issue embedded in the existing approaches. Especially, we theoretically prove that the merits of the proposal boil down to guaranteed estimator variance and bias reduction. Code is available: at https://github.com/ssl-codelab/uota.

Fairness is a principal social value that can be observed in civilisations around the world. A manifestation of this is in social agreements, often described in texts, such as contracts. Yet, despite the prevalence of such, a fairness metric for texts describing a social act remains wanting. To address this, we take a step back to consider the problem based on first principals. Instead of using rules or templates, we utilise social psychology literature to determine the principal factors that humans use when making a fairness assessment. We then attempt to digitise these using word embeddings into a multi-dimensioned sentence level fairness perceptions vector to serve as an approximation for these fairness perceptions. The method leverages a pro-social bias within word embeddings, for which we obtain an F1= 81.0. A second approach, using PCA and ML based on the said fairness approximation vector produces an F1 score of 86.2. We detail improvements that can be made in the methodology to incorporate the projection of sentence embedding on to a subspace representation of fairness.

Although artificial intelligence has been the subject of academic research since the 1950s and has been used commercially in some industries for decades, it is still in its infancy across much of the broader economy. The rapid adoption of this technology, along with the unique privacy, security and liability issues associated with it, has created opportunities for lawyers to help their clients capture its economic value while ensuring its use is ethical and legal. However, before advising clients on AI issues, lawyers should have some basic technical knowledge to answer questions about legal compliance. Machine learning algorithms are incredibly complex, learning billions of rules from datasets and applying those rules to arrive at an output recommendation. Even the most precise and well-designed AI systems are probabilistic in nature, guaranteeing that the system will, at some point, produce an incorrect result.

Typically, non-governmental organizations collect and curate large bodies of human rights information, with the goal of making these collections useful for advocates. Manually processing these documents can take several days, particularly when they're published in unfamiliar languages or in PDF format which is difficult to search through. As a result, many NGOs face a large backlog of documents that remain to be processed, and by the time they're added to collections new documentation often supersedes them. Based in Geneva, HURIDOCS has been developing tools to manage and analyze collections of human rights evidence, law and research for nearly four decades. In 2016, they had an idea: What if machine learning could skim through documents, make terms extractable, and classify the content to catalog documents more quickly?

"His companies have faced allegations of sexual harassment and poor working conditions; in October, a federal jury ordered Tesla to pay $137 million to a Black employee who accused the automaker of ignoring racial abuse. The businesses have also been fined for numerous regulatory violations. The feds are probing Tesla's Autopilot software, which has been involved in an alarming number of crashes with parked emergency vehicles, resulting in injuries and death. The company's expansion in China required cozying up to its repressive autocrats." So naturally, let's make him Time's Person of the Year! No one can accuse Time of rolling over for its pick.

Coinbase has built the world's leading compliant cryptocurrency platform serving over 73 million accounts in more than 100 countries. With multiple successful products, and our vocal advocacy for blockchain technology, we have played a major part in mainstream awareness and adoption of cryptocurrency. We are proud to offer an entire suite of products that are helping build the cryptoeconomy and increase economic freedom around the world. There are a few things we look for across all hires we make at Coinbase, regardless of role or team. First, we look for signals that a candidate will thrive in a culture like ours, where we default to trust, embrace feedback, disrupt ourselves, and expect sustained high performance because we play as a championship team.

The researchers pored through nearly 50,000 medical records and found that the software had recommended Black patients for additional care about half the time they should have, while white patients were recommended for additional care at a far higher rate. The reason, Gray explained, was that the algorithms factored in medical histories in predicting how much each patient was likely to cost the health care system if left untreated. This meant that white patients, who typically have better access to health care due to a variety of factors rooted in systemic racism, were given priority for certain treatments.

China has warned the United States that it would "strike back" in response to any "reckless" actions, urging Washington to withdraw its recent passing of sanctions targeting people and entities tied to human rights abuses committed by Beijing. The United States imposed sweeping human rights-related sanctions on Friday against Chinese individuals and entities, adding individuals and entities tied to Myanmar, North Korea and Bangladesh. China's Foreign Ministry spokesperson Wang Wenbin denounced the sanctions as "perverse actions". "We urge the US to immediately withdraw the relevant wrong decision and stop interfering in China's internal affairs and harming China's interests. "If the US acts recklessly, China will take effective measures to strike back resolutely," Wang said during a news conference in Beijing on Monday.

New York City will become the latest jurisdiction to regulate an employer's use of artificial intelligence (AI) and other "automated employment decision tools" in screening job candidates. The law is intended to curb bias in hiring and promotion decisions. Effective January 2, 2023, employers and employment agencies will be prohibited from using automated decision tools for screening employment or promotion candidates unless: (1) the tool has undergone an independent bias audit no more than one year prior to its use; and (2) certain information relating to the audit results is made publicly available on the employer's or employment agency's website. Additionally, companies will be required to notify employees or job applicants whether an AI tool was used to make employment decisions (amongst other notice requirements). NYC's law comes on the heels of the Equal Employment Opportunity Commission's (EEOC) announcement in October 2021 of an initiative to examine how AI technology is "fundamentally changing" the way employment decisions are made.

Discrete and continuous manufacturing lines generate a high volume of products at low latency, ranging from milliseconds to a few seconds. To identify defects at the same throughput of production, camera streams of images must be processed at low latency. Additionally, factories may have low network bandwidth or intermittent cloud connectivity. In such scenarios, you may need to run the defect detection system on your on-premises compute infrastructure, and upload the processed results for further development and monitoring purposes to the AWS Cloud. This hybrid approach with both local edge hardware and the cloud can address the low latency requirements and help reduce storage and network transfer costs to the cloud.