Fostering Innovation with Generative AI: A Study on Human-AI Collaborative Ideation and User Anonymity

Collaborative ideation is a key aspect of many innovation processes. However, a lack of proper support can hinder the process and limit the ability of participants to generate innovative ideas. Thus, we introduce AI-deation, a digital environment for collaborative ideation. At the heart of the system is an AI collaborator powered by Generative Artificial Intelligence that participates in the ideation process by automatically suggesting new ideas. Moreover, the submitted ideas are processed by a Large Language Model, acting as an idea editor. It strengthens the anonymity of the contributions to alleviate fears of judgment. We studied this system in a randomized experiment where groups solved complex problems in two conditions: humans only and humans supported by an AI collaborator. Results show that the idea editor effectively strengthened the participants' anonymity. Despite being more innovative, the AI collaborator did not significantly influence the participants' innovativeness.

Guiding Language Models of Code with Global Context using Monitors

Language models of code (LMs) work well when the surrounding code provides sufficient context. This is not true when it becomes necessary to use types, functionality or APIs defined elsewhere in the repository or a linked library, especially those not seen during training. LMs suffer from limited awareness of such global context and end up hallucinating. Integrated development environments (IDEs) assist developers in understanding repository context using static analysis. We extend this assistance, enjoyed by developers, to LMs. We propose monitor-guided decoding (MGD) where a monitor uses static analysis to guide the decoding. We construct a repository-level dataset PragmaticCode for method-completion in Java and evaluate MGD on it. On models of varying parameter scale, by monitoring for type-consistent object dereferences, MGD consistently improves compilation rates and agreement with ground truth. Further, LMs with fewer parameters, when augmented with MGD, can outperform larger LMs. With MGD, SantaCoder-1.1B achieves better compilation rate and next-identifier match than the much larger text-davinci-003 model. We also conduct a generalizability study to evaluate the ability of MGD to generalize to multiple programming languages (Java, C# and Rust), coding scenarios (e.g., correct number of arguments to method calls), and to enforce richer semantic constraints (e.g., stateful API protocols). Our data and implementation are available at https://github.com/microsoft/monitors4codegen .

Tutoriel – Utiliser ReaderBench pour analyser automatiquement des discussions — Recherche en éducation, Inspé, Univ. Grenoble Alpes