This chapter discusses several key aspects and basics of the European Convention on Human Rights. First, it explains two main principles underlying the ECHR system as a whole: the principle of effective protection of fundamental rights and the principle of subsidiarity. In addition, the double role of the Court within the Convention system is set out, which is to offer individual redress or individual justice when Convention rights have been violated on the national level, and to clarify Convention rights standards (which is the Court’s constitutional role). The Chapter further addresses the three main stages of the Court’s review, which correspond to the structure of most of the Convention rights: (1) deciding on the applicability and interpretation of a Convention right; (2) determining if there is an interference with the right; (3) reviewing whether a justification can be given for this interference. Finally, a typology is provided of Convention rights according to the possibilities for restricting these rights, e.g. paying attention to absolute, non-derobable rights and distinguishing between different types derogable rights.

A reinforcement learning (RL) agent acts in an environment, observing its state and receiving rewards. From its experience of a stream of acting then observing the resulting state and reward, it must determine what to do given its goal of maximizing accumulated reward. This chapter considers fully observable (page 29), single-agent reinforcement learning.

This Chapter first explains what the ECtHR regards as the object of review in cases where an alleged violation of the Convention is caused by legislation: is this the legislation as such (which would invite general and more abstract review), or is it the individual decision applying this legislation (which would invite individualised and more concrete review), or perhaps both? The chapter then turns to discussing how the Court has tried to reconcile its task of offering individual justice and general constitutional interpretations. Specific attention is paid in this regard to the role of precedent-based reasoning in the Court’s case law and to case-based review, incrementalism and the development of general principles. Finally, the legal effect of these general principles is discussed (so-called ’res interpretata’ or force of interpretation), in contrast to the application thereof to the facts of the individual case.

This chapter considers simple forms of reasoning in terms of propositions – statements that can be true or false. Some reasoning includes model finding, finding logical consequences, and various forms of hypothetical reasoning. Semantics forms the foundations of specification of facts, reasoning, and debugging.

In many, if not most cases, the ECtHR relies on its own precedents in interpreting the Convention, but some cases may present the Court with new and difficult questions of interpretation. To answer these, the Court relies on a number of specific principles and methods of interpretation. The Court is guided in its work by several core principles, such as effectiveness, evolutive interpretation and autonomous interpretatio, but these are still relatively abstract in nature and they may not suffice to answer a concrete question of interpretation. For that reason, the Court often also relies on the methods of interpretation as described in the provisions of the Vienna Convention on the Law of Treaties, such as textual interpretation, interpretation in light of the travaux préparatoires and internally harmonising interpretation. This chapter discusses the Court’s use of these three methods. In addition to these methods, the Court has opted for a particular refinement of one of the Vienna Convention’s methods, which is consensus or common ground interpretation. This method and the various sources for the Court’s finding of a consensus are also discussed in this chapter.

Learning is the ability of an agent to improve its behavior based on experience. This could mean the following: • The range of behaviors is expanded; the agent can do more. • The accuracy on tasks is improved; the agent can do things better. • The speed is improved; the agent can do things faster.

Deterministic planning is the process of finding a sequence of actions to achieve a goal. Because an agent does not usually achieve its goals in one step, what it should do at any time depends on what it will do in the future. What it will do in the future depends on the state it is in, which, in turn, depends on what it has done in the past. This chapter presents representations of actions and their effects, and some offline algorithms for an agent to find a plan to achieve its goals from a given state.