This paper introduces a novel benchmark that has been designed as a test bed for evaluating whether artificial agents are able to understand how to perform everyday activities, with a focus on the cooking domain. Understanding how to cook recipes is a highly challenging endeavour due to the underspecified and grounded nature of recipe texts, combined with the fact that recipe execution is a knowledge-intensive and precise activity. The benchmark comprises a corpus of recipes, a procedural semantic representation language of cooking actions, qualitative and quantitative kitchen simulators, and a standardised evaluation procedure. Concretely, the benchmark task consists in mapping a recipe formulated in natural language to a set of cooking actions that is precise enough to be executed in the simulated kitchen and yields the desired dish. To overcome the challenges inherent to recipe execution, this mapping process needs to incorporate reasoning over the recipe text, the state of the simulated kitchen environment, common-sense knowledge, knowledge of the cooking domain, and the action space of a virtual or robotic chef. This benchmark thereby addresses the growing interest in human-centric systems that combine natural language processing and situated reasoning to perform everyday activities.

One of AI’s grand challenges consists in the development of autonomous agents with communication systems offering the robustness, flexibility and adaptivity found in human languages. While the processes through which children acquire language are by now relatively well understood, a faithful computational operationalisation of the underlying mechanisms is still lacking. Two main cognitive processes are involved in child language acquisition. First, children need to reconstruct the intended meaning of observed utterances, a process called intention reading. Then, they can gradually abstract away from concrete utterances in a process called pattern finding and acquire productive schemata that generalise over form and meaning. In this paper, we introduce a mechanistic model of the intention reading process and its integration with pattern finding capacities. Concretely, we present an agent-based simulation in which an agent learns a grammar that enables them to ask and answer questions about a scene. This involves the reconstruction of queries that correspond to observed questions based on the answer and scene alone, and the generalization of linguistic schemata based on these reconstructed question-query pairs. The result is a productive grammar which can be used to map between natural language questions and queries without ever having observed the queries.