Armando Solar-Lezama

Neurosymbolic programming is an emerging area that bridges the areas of deep learning and program synthesis. As in classical machine learning, the goal is to learn functions from data. However, these functions are represented as programs that can use neural modules in addition to symbolic primitives and are induced using a combination of symbolic search and gradient-based optimization. Neurosymbolic programming can offer multiple advantages over end-to-end deep learning. Programs can sometimes naturally represent long-horizon, procedural tasks that are difficult to perform using deep networks. Neurosymbolic representations are also, commonly, easier to interpret and formally verify than neural networks. The restrictions of a programming language can serve as a form of regularization and lead to more generalizable and data-efficient learning. Compositional programming abstractions can also be a natural way of reusing learned modules across learning tasks. In this monograph, the authors illustrate these potential benefits with concrete examples from recent work on neurosymbolic programming. They also categorize the main ways in which symbolic and neural learning techniques come together in this area and conclude with a discussion of the open technical challenges in the field. The comprehensive review of neurosymbolic programming introduces the reader to the topic and provides an insightful treatise on an increasingly important topic at the intersection of programming languages and machine learning.

Database management systems (DBMS) typically provide an application programming interface for users to issue queries using query languages such as SQL. Many such languages were originally designed for business data processing applications. While these applications are still relevant, two other classes of applications have become important users of data management systems: (1) web applications that issue queries programmatically to the DBMS, and (2) data analytics involving complex queries that allow data scientists to better understand their datasets. Unfortunately, existing query languages provided by database management systems are often far from ideal for these application domains. Computer-Assisted Query Formulation describes a set of technologies that assist users in specifying database queries for different application domains. The goal of such systems is to bridge the gap between current query interfaces provided by database management systems and the needs of different usage scenarios that are not well served by existing query languages. This monograph discusses the different interaction modes that such systems provide and the algorithms used to infer user queries. In particular, it focuses on a new class of systems built using program synthesis techniques, and furthermore discusses opportunities in combining synthesis and other methods used in prior systems to infer user queries.