The goal of this seminar is to exchange recent research results and foster novel ideas on topics at the frontier between Abstract Interpretation and Constraint Programming. Although, a priori, these two fields do not share the same methods nor goals, connections between these fields exist and have recently seen an increase in interest. A natural intersection point of these communities lies in the analysis and verification of numeric-heavy programs and systems, over continuous domains and floating-point computations, with application to verifying the safety of critical embedded software. The objective of the seminar is to discuss recent advances and ongoing research on these connections, which promise: a deeper understanding of the theoretical connections between AI and CP; an improvement of AI and CP methods by importing, in each field, results and methods from the other field; the design of hybrid and unified AI-CP methods. Technical topics of interest include: fixpoint computation; numeric abstract domains; linear and non-linear abstractions; reduction techniques; combinations of discrete and continuous constraints; applications to the analysis of embedded systems, hybrid systems, and cyber-physical systems; floating-point constraints; robustness analysis and analysis of rounding errors; algebraic formalisation of CP techniques; implementation and applications. Very few researchers are currently expert in both fields, which limits our ability to unveil their connections and limits the opportunity for effective combinations. The seminar will gather expert researchers from both communities that have been working, from either side, to close the gap. The seminar is organized in the scope of the ANR Project Coverif which studies theoretical and practical aspects of combinations of AI and CP.