Reinforcement learning is well suited for optimizing policies of recommender systems.

We study fair machine learning (ML) under predictive uncertainty to enable reliable and trustworthy decision-making. The seminal work of'equalized coverage' proposed an uncertainty-aware fairness notion. However, it does not guarantee equal coverage rates across more fine-grained groups (e.g., low-income females) conditioning on the true label and is biased in the assessment of uncertainty. To tackle these limitations, we propose a new uncertainty-aware fairness -- Equal Opportunity of Coverage (EOC) -- that aims to achieve two properties: (1) coverage rates for different groups with similar outcomes are close, and (2) the coverage rate for the entire population remains at a predetermined level. Further, the prediction intervals should be narrow to be informative. We propose Binned Fair Quantile Regression (BFQR), a distribution-free post-processing method to improve EOC with reasonable width for any trained ML models. It first calibrates a hold-out set to bound deviation from EOC, then leverages conformal prediction to maintain EOC on a test set, meanwhile optimizing prediction interval width. Experimental results demonstrate the effectiveness of our method in improving EOC.

Early research on AI typically involved qualitative demonstrations of intelligent behavior, with novelty being the primary focus. However, as the field has matured, there have been increasing demands for more careful evaluation using quantitative measures of behavior. In some cases, the response has taken the guise of formal analyses, and in others, it has emphasized comparisons between system and human behavior, but the predominant movement has been toward empirical studies of AI methods. As a result, techniques for experimental design, exploratory data analysis, and statistical testing, originally developed in other fields, have become increasingly relevant for AI researchers. Paul Cohen's book Empirical Methods for Artificial Intelligence aims to encourage this trend by providing AI practitioners with the knowledge and tools needed for careful empirical evaluation.

The insurance sector is one of the most old-fashioned and resistant to change -- so artificial intelligence will have an even greater effect. The insurance sector is one of the most old-fashioned and resistant to change, and this is why AI will have a greater impact on that with respect to more receptive industries. The collection of data of new types (i.e., unstructured data such as reports, images, contracts, etc.) and the use of new algorithms are disrupting the sector in several ways. This is a really simplistic representation of the insurance business in the last fifty years, and I am aware that insurance experts might disagree with me in many different ways. There are a couple of further features to be pointed out: first of all, insurance has historically been sold not bought, which means that brokers and agents were essential to tracking new customers and to even retain old ones.

LinkedIn's knowledge graph is a large knowledge base built upon "entities" on LinkedIn, such as members, jobs, titles, skills, companies, geographical locations, schools, etc. To solve the challenges we face when building the LinkedIn knowledge graph, we apply machine learning techniques, which is essentially a process of data standardization on user-generated content and external data sources, in which machine learning is applied to entity taxonomy construction, entity relationship inference, data representation for downstream data consumers, insight extraction from graph, and interactive data acquisition from users to validate our inference and collect training data. By mining member profiles for entity candidates and utilizing external data sources and human validations to enrich candidate attributes, we created tens of thousands of skills, titles, geographical locations, companies, certificates, etc., to which we can map members. Given the power-law nature of the member coverage of entities, linguistic experts at LinkedIn manually translate the top entities with high member coverages into international languages to achieve high precision, and PSCFG-based machine translation models are applied to automatically translate long-tail entities to achieve high recall.

For example, a new customer support ticket with email threads between a customer and a CSR arrive and the machine learning model would predict a categorization and tell you how confident it was about that particular prediction. The media coverage seems to imply that AI is only the domain of the technology elite – companies such as Amazon Apple, Facebook, Google, IBM, Microsoft, Salesforce, Tesla, Uber – who can afford to assemble large teams of machine learning experts and invest $100M. But if a business did this without thinking about how it would also get high quality, high volume customized training data from which the machine learning model could learn you would have a mismatch between expectations ("we have a great algorithm") to outcome ("our model is only 60% accurate"). So over time the model can handle an increasing percentage of the customer support ticket classification work and the business can greatly increase the volume of tickets it classifies.

This is the final part in a series where we use machine learning and natural language processing to analyze articles published in tech news sites in order to gain insights about the state of the startup industry. Let's visualize the coverage per industry for all the articles published in the last ten years to find out: The most popular startup industry in the last ten years has been Mobile. Right when this industry started to gain visibility in 2013, Oculus Rift was the top keyword by a wide margin. The only way to perform an analysis like this is using machine learning and natural language processing, since there's no way we can get a person to read through and interpret 270,000 articles.

The central challenge in commonsense knowledge representation research is to develop content theories that achieve a high degree of both competency and coverage. We describe a new methodology for constructing formal theories in commonsense knowledge domains that complements traditional knowledge representation approaches by first addressing issues of coverage. These concepts are sorted into a manageable number of coherent domains, one of which is the representational area of commonsense human memory. These representational areas are then analyzed using more traditional knowledge representation techniques, as demonstrated in this article by our treatment of commonsense human memory.