In domains ranging from computer vision to natural language processing, machine learning models have been shown to exhibit stark disparities, often performing worse for members of traditionally underserved groups. One factor contributing to these performance gaps is a lack of representation in the data the models are trained on. It is often unclear, however, how to operationalize representativeness in specific applications. Here we formalize the problem of creating equitable training datasets, and propose a statistical framework for addressing this problem. We consider a setting where a model builder must decide how to allocate a fixed data collection budget to gather training data from different subgroups. We then frame dataset creation as a constrained optimization problem, in which one maximizes a function of group-specific performance metrics based on (estimated) group-specific learning rates and costs per sample. This flexible approach incorporates preferences of model-builders and other stakeholders, as well as the statistical properties of the learning task. When data collection decisions are made sequentially, we show that under certain conditions this optimization problem can be efficiently solved even without prior knowledge of the learning rates. To illustrate our approach, we conduct a simulation study of polygenic risk scores on synthetic genomic data -- an application domain that often suffers from non-representative data collection. We find that our adaptive sampling strategy outperforms several common data collection heuristics, including equal and proportional sampling, demonstrating the value of strategic dataset design for building equitable models.

In populous countries, pending legal cases have been growing exponentially. There is a need for developing techniques for processing and organizing legal documents. In this paper, we introduce a new corpus for structuring legal documents. In particular, we introduce a corpus of legal judgment documents in English that are segmented into topical and coherent parts. Each of these parts is annotated with a label coming from a list of pre-defined Rhetorical Roles. We develop baseline models for automatically predicting rhetorical roles in a legal document based on the annotated corpus. Further, we show the application of rhetorical roles to improve performance on the tasks of summarization and legal judgment prediction. We release the corpus and baseline model code along with the paper.

Classification and topic modeling are popular techniques in machine learning that extract information from large-scale datasets. By incorporating a priori information such as labels or important features, methods have been developed to perform classification and topic modeling tasks; however, most methods that can perform both do not allow for guidance of the topics or features. In this paper, we propose a method, namely Guided Semi-Supervised Non-negative Matrix Factorization (GSSNMF), that performs both classification and topic modeling by incorporating supervision from both pre-assigned document class labels and user-designed seed words. We test the performance of this method through its application to legal documents provided by the California Innocence Project, a nonprofit that works to free innocent convicted persons and reform the justice system. The results show that our proposed method improves both classification accuracy and topic coherence in comparison to past methods like Semi-Supervised Non-negative Matrix Factorization (SSNMF) and Guided Non-negative Matrix Factorization (Guided NMF).

Precision medicine relies on quick and accurate detection of patients with severe and heterogeneous illnesses. For example, machine learning can be used to identify leukaemia patients based on their blood transcriptomes. However, due to privacy regulations, there is a growing gap between what is permitted and what is technically feasible. Introduction of Swarm Learning-a decentralised machine learning approach combining edge computing, peer-to-peer-networking, and coordination while preserving confidentiality, allows integration of any medical data from anywhere in the world, accelerating the adoption of precision medicine.

Providers of services that manipulate images and videos will be required to verify user identities and respect social morality.

A whistleblower who exposed illegal working conditions in a factory making Amazon's Alexa devices says he was tortured before being jailed by Chinese authorities. Tang Mingfang, 43, was jailed after he revealed how the Foxconn factory in the southern Chinese city of Hengyang used schoolchildren working illegally long hours to manufacture Amazon's popular Echo, Echo Dot and Kindle devices. Now, after spending two years in prison, he is appealing to the higher courts to clear his name. He has taken the difficult decision to talk publicly, despite being aware of the risks of reprisals, because he believes Amazon and its founder, Jeff Bezos, have a responsibility to support his appeal and that the Observer also has a responsibility to highlight his case. Tang, who is married with a nine-year-old son, said his father – who died while he was in prison – would have wanted him to speak up when he saw young workers being abused.

With mission-critical operations, artificial intelligence (AI) has the potential to produce incredible benefits – not only for businesses but also for the people they serve and employ. You see it when systems detect fraudulent purchases and keep a consumer's account safe. It's in autonomous and self-driving cars, which are programmed to help keep drivers safe and avoid collisions. In each of these examples, AI is a tool to learn complex patterns – including some that are practically undetectable. The result is more impactful and, with appropriate oversight, better and fairer decision-making.

As shown in Figure 1, around half of global supply chain executives are pressured by regulators, company executives, end users, etc. to improve their supply chain sustainability. Consequently, 59% of enterprises invested in improving the sustainability of their supply chain. A sustainable supply chain is an important part of improving a company's environmental, social, and governance (ESG) standards which have an impact on attracting more customers and investors. It is responsible for the bulk of scope 3 GHG emissions of a company. Additionally, corporations that source raw materials or intermediate items from developing nations could unintentionally abuse their suppliers' employees who work in inhumane conditions.

Ranking, recommendation, and retrieval systems are widely used in online platforms and other societal systems, including e-commerce, media-streaming, admissions, gig platforms, and hiring. In the recent past, a large "fair ranking" research literature has been developed around making these systems fair to the individuals, providers, or content that are being ranked. Most of this literature defines fairness for a single instance of retrieval, or as a simple additive notion for multiple instances of retrievals over time. This work provides a critical overview of this literature, detailing the often context-specific concerns that such an approach misses: the gap between high ranking placements and true provider utility, spillovers and compounding effects over time, induced strategic incentives, and the effect of statistical uncertainty. We then provide a path forward for a more holistic and impact-oriented fair ranking research agenda, including methodological lessons from other fields and the role of the broader stakeholder community in overcoming data bottlenecks and designing effective regulatory environments.

Fox News Flash top headlines are here. Check out what's clicking on Foxnews.com. The Polk County Sheriff's Office in Florida on Thursday announced charges against 68 suspected drug dealers as part of an undercover operation using social media and dating apps. The six-month operation dubbed "Swipe Left for Meth" -- a play on Grindr and other dating apps that require users to "swipe" through scores of potential dates in their area -- concluded in the arrests of 60 individuals and securement of eight arrest warrants for individuals still at-large related to drug sales or possession. "We've known for some time that suspects will use the internet and social media to prey upon children online, or to engage in prostitution, but this is something we are seeing more and more of in Polk County -- suspects who are using dating apps to sell illegal narcotics," Polk County Sheriff Grady Judd said in a statement." Suspects are getting more creative, but so are our detectives."