Date:
Location:
Speaker: Benjamin Laufer (Cornell Tech)
Talk Title: Fine-tuning Games
Abstract: Major advances in Machine Learning (ML) and Artificial Intelligence (AI) increasingly take the form of developing and releasing general-purpose models. These models are designed to be adapted by other businesses and agencies to perform a particular, domain-specific function. This process has become known as adaptation or fine-tuning. This talk will offer a model of the fine-tuning process where a Generalist brings the technological product (here an ML model) to a certain level of performance, and one or more Domain-specialist(s) adapts it for use in a particular domain. Both entities are profit-seeking and incur costs when they invest in the technology, and they must reach a bargaining agreement on how to share the revenue for the technology to reach the market. For a relatively general class of cost and revenue functions, we characterize the conditions under which the fine-tuning game yields a profit-sharing solution. We observe that any potential domain-specialization will either "contribute," "free-ride," or "abstain" in their uptake of the technology, and we provide conditions yielding these different strategies. We show how methods based on bargaining solutions and sub-game perfect equilibria provide insights into the strategic behavior of firms in these types of interactions, and we find that profit-sharing can still arise even when one firm has significantly higher costs than another. We also provide methods for identifying Pareto-optimal bargaining arrangements for a general set of utility functions. Finally, we will discuss some extensions of this model to questions about regulation, liability, and harms in AI systems.
Shorter description: In this work, we provide a model of the strategic dynamics implicit in domain adaptation of a general-purpose AI/ML technology. We do some analysis suggesting whether/which/how domains will adapt a general-purpose model. For example, depending on the cost of investment, a domain might either "abstain," "contribute," or "free-ride." We will discuss current findings as well as present work extending the model to questions around regulation, liability, and harms in AI systems.