🔬 Research Summary by Jayanth Yetukuri, a final year Ph.D. student at UCSC, advised by Professor Yang Liu, where his research focuses on improving the trustworthiness of Machine Learning models.
[Original paper by Jayanth Yetukuri, Ian Hardy, and Yang Liu]
Overview: Recent years have seen a proliferation of Machine Learning systems in several critical decision-making domains. Actionable Recourse provides necessary action to an adversely affected individual by the model’s decision. This paper focuses on allowing such individuals to maneuver the recourse generation process by capturing their individual preferences.
Introduction
Actionable Recourse is a list of actions an individual can take to obtain a desired outcome from a fixed Machine Learning model. Several domains such as lending, insurance, resource allocation, and hiring decisions are required to suggest recourses to ensure the trust of a decision system; in such scenarios, it is critical to ensure the actionability (the viability of taking a suggested action) of recourse, otherwise, the suggestions are pointless.
Existing research focuses on providing feasible recourses, yet comprehensive literature on understanding and incorporating user preferences within the recourse generation mechanism is lacking. Efforts to elicit user preferences include recent work by De Toni et al. 2022. The authors provide an interactive human-in-the-loop approach, where a user continuously interacts with the system. However, learning user preferences by asking them to select from one of the partial interventions provided is a derivative of providing a diverse set of recourse candidates.
We argue that the inherent problem of feasibility can be solved more accurately by capturing and understanding Alice’s recourse preference and adhering to her constraints, which can vary between Hard Rules, such as being unable to bring a co-applicant, and Soft Rules, such as hesitation to reduce the amount, which should not be interpreted as unwillingness.
Key Insights
Motivated by the above considerations, we capture soft user preferences and hard constraints and identify recourse based on local desires without affecting the success rate of identifying recourse. For example, consider Alice prefers to have 80% of the recourse “cost” from loan duration and only 20% from the loan amount, meaning she prefers to have recourse with a minor reduction in the loan amount. Such recourse enables Alice to get the benefits of a loan on her terms and can easily be calculated according to Alice’s desire. Hence, user-preferred recourse is obtained by solving a custom optimization for individual preferences.
User preferences can be captured via soft constraints in three simple forms:
i) scoring continuous features,
ii) bounding feature values, and
iii) ranking categorical features.
These preferences can be embedded into the gradient-based recourse identification approach approach to design User Preferred Actionable Recourse (UP-AR). User Preferred Actionable Recourse (UP-AR) consists of two stages. The first stage generates a candidate recourse by following a connected gradient-based iterative approach. The second stage then improves upon the redundancy metric of the generated recourse for better actionability.
UP-AR holistically performs favorably to its counterparts. Critically, it respects feature constraints (which are fundamental to actionable recourse) while maintaining a significantly low redundancy and sparsity. This indicates that it tends to change fewer necessary features. Its speed makes it tractable for real-world use, while its proximity values show that it recovers relatively low-cost recourse. These results highlight the promise of UP-AR as a performative, low-cost option for calculating recourse when user preferences are paramount. UP-AR shows consistent improvements over all the performance metrics.
Between the lines
In this study, we propose to capture different forms of user preferences and propose an optimization function to generate actionable recourse adhering to such constraints. We further provide an approach to generate a connected recourse guided by the user. We show how UP-AR adheres to soft constraints by evaluating user satisfaction in fractional cost ratio. We emphasize the need to capture various user preferences and communicate with the user in a comprehensible form. This work motivates further research on how truthful reporting of preferences can help improve overall user satisfaction.