Counterfactual Token Generation in Large Language Models

"Sure, I am happy to generate a story for you: Captain Lyra stood at the helm of her trusty ship, the Maelstrom's Fury, gazing out at the endless sea. [...] Lyra's eyes welled up with tears as she realized the bitter truth - she had sacrificed everything for fleeting riches, and lost the love of her crew, her family, and herself." Although this story, generated by a large language model, is captivating, one may wonder -- how would the story have unfolded if the model had chosen "Captain Maeve" as the protagonist instead? We cannot know. State-of-the-art large language models are stateless -- they maintain no internal memory or state. Given a prompt, they generate a sequence of tokens as an output using an autoregressive process. As a consequence, they cannot reason about counterfactual alternatives to tokens they have generated in the past. In this work, our goal is to enhance them with this functionality. To this end, we develop a causal model of token generation that builds upon the Gumbel-Max structural causal model. Our model allows any large language model to perform counterfactual token generation at almost no cost in comparison with vanilla token generation, it is embarrassingly simple to implement, and it does not require any fine-tuning nor prompt engineering. We implement our model on Llama 3 8B-instruct and conduct both qualitative and quantitative analyses of counterfactually generated text. We conclude with a demonstrative application of counterfactual token generation for bias detection, unveiling interesting insights about the model of the world constructed by large language models.

Finding Counterfactually Optimal Action Sequences in Continuous State Spaces

Humans performing tasks that involve taking a series of multiple dependent actions over time often learn from experience by reflecting on specific cases and points in time, where different actions could have led to significantly better outcomes. While recent machine learning methods to retrospectively analyze sequential decision making processes promise to aid decision makers in identifying such cases, they have focused on environments with finitely many discrete states. However, in many practical applications, the state of the environment is inherently continuous in nature. In this paper, we aim to fill this gap. We start by formally characterizing a sequence of discrete actions and continuous states using finite horizon Markov decision processes and a broad class of bijective structural causal models. Building upon this characterization, we formalize the problem of finding counterfactually optimal action sequences and show that, in general, we cannot expect to solve it in polynomial time. Then, we develop a search method based on the $A^*$ algorithm that, under a natural form of Lipschitz continuity of the environment's dynamics, is guaranteed to return the optimal solution to the problem. Experiments on real clinical data show that our method is very efficient in practice, and it has the potential to offer interesting insights for sequential decision making tasks.

On the Within-Group Discrimination of Screening Classifiers

Screening classifiers are increasingly used to identify qualified candidates in a variety of selection processes. In this context, it has been recently shown that, if a classifier is calibrated, one can identify the smallest set of candidates which contains, in expectation, a desired number of qualified candidates using a threshold decision rule. This lends support to focusing on calibration as the only requirement for screening classifiers. In this paper, we argue that screening policies that use calibrated classifiers may suffer from an understudied type of within-group discrimination -- they may discriminate against qualified members within demographic groups of interest. Further, we argue that this type of discrimination can be avoided if classifiers satisfy within-group monotonicity, a natural monotonicity property within each of the groups. Then, we introduce an efficient post-processing algorithm based on dynamic programming to minimally modify a given calibrated classifier so that its probability estimates satisfy within-group monotonicity. We validate our algorithm using US Census survey data and show that within-group monotonicity can be often achieved at a small cost in terms of prediction granularity and shortlist size.

Counterfactual Explanations in Sequential Decision Making Under Uncertainty

Methods to find counterfactual explanations have predominantly focused on one step decision making processes. In this work, we initiate the development of methods to find counterfactual explanations for decision making processes in which multiple, dependent actions are taken sequentially over time. We start by formally characterizing a sequence of actions and states using finite horizon Markov decision processes and the Gumbel-Max structural causal model. Building upon this characterization, we formally state the problem of finding counterfactual explanations for sequential decision making processes. In our problem formulation, the counterfactual explanation specifies an alternative sequence of actions differing in at most k actions from the observed sequence that could have led the observed process realization to a better outcome. Then, we introduce a polynomial time algorithm based on dynamic programming to build a counterfactual policy that is guaranteed to always provide the optimal counterfactual explanation on every possible realization of the counterfactual environment dynamics. We validate our algorithm using both synthetic and real data from cognitive behavioral therapy and show that the counterfactual explanations our algorithm finds can provide valuable insights to enhance sequential decision making under uncertainty.

Group Testing under Superspreading Dynamics

Testing is recommended for all close contacts of confirmed COVID-19 patients. However, existing group testing methods are oblivious to the circumstances of contagion provided by contact tracing. Here, we build upon a well-known semi-adaptive pool testing method, Dorfman's method with imperfect tests, and derive a simple group testing method based on dynamic programming that is specifically designed to use the information provided by contact tracing. Experiments using a variety of reproduction numbers and dispersion levels, including those estimated in the context of the COVID-19 pandemic, show that the pools found using our method result in a significantly lower number of tests than those found using standard Dorfman's method, especially when the number of contacts of an infected individual is small. Moreover, our results show that our method can be more beneficial when the secondary infections are highly overdispersed.

Bridging Machine Learning and Mechanism Design towards Algorithmic Fairness

Decision-making systems increasingly orchestrate our world: how to intervene on the algorithmic components to build fair and equitable systems is therefore a question of utmost importance; one that is substantially complicated by the context-dependent nature of fairness and discrimination. Modern systems incorporate machine-learned predictions in broader decision-making pipelines, implicating concerns like constrained allocation and strategic behavior that are typically thought of as mechanism design problems. Although both machine learning and mechanism design have individually developed frameworks for addressing issues of fairness and equity, in some complex decision-making systems, neither framework is individually sufficient. In this paper, we develop the position that building fair decision-making systems requires overcoming these limitations which, we argue, are inherent to the individual frameworks of machine learning and mechanism design. Our ultimate objective is to build an encompassing framework that cohesively bridges the individual frameworks. We begin to lay the ground work towards achieving this goal by comparing the perspective each individual discipline takes on fair decision-making, teasing out the lessons each field has taught and can teach the other, and highlighting application domains that require a strong collaboration between these disciplines.

A Spatiotemporal Epidemic Model to Quantify the Effects of Contact Tracing, Testing, and Containment

Motivated by the current COVID-19 outbreak, we introduce a novel epidemic model based on marked temporal point processes that is specifically designed to make fine-grained spatiotemporal predictions about the course of the disease in a population. Our model can make use and benefit from data gathered by a variety of contact tracing technologies and it can quantify the effects that different testing and tracing strategies, social distancing measures, and business restrictions may have on the course of the disease. Building on our model, we use Bayesian optimization to estimate the risk of exposure of each individual at the sites they visit, the percentage of symptomatic individuals, and the difference in transmission rate between asymptomatic and symptomatic individuals from historical longitudinal testing data. Experiments using real COVID-19 data and mobility patterns from T\"{u}bingen, a town in the southwest of Germany, demonstrate that our model can be used to quantify the effects of tracing, testing, and containment strategies at an unprecedented spatiotemporal resolution. To facilitate research and informed policy-making, particularly in the context of the current COVID-19 outbreak, we are releasing an open-source implementation of our framework at https://github.com/covid19-model.

Decisions, Counterfactual Explanations and Strategic Behavior

Data-driven predictive models are increasingly used to inform decisions that hold important consequences for individuals and society. As a result, decision makers are often obliged, even legally required, to provide explanations about their decisions. In this context, it has been increasingly argued that these explanations should help individuals understand what would have to change for these decisions to be beneficial ones. However, there has been little discussion on the possibility that individuals may use the above counterfactual explanations to invest effort strategically in order to maximize their chances of receiving a beneficial decision. In this paper, our goal is to find policies and counterfactual explanations that are optimal in terms of utility in such a strategic setting. To this end, we first show that, given a pre-defined policy, the problem of finding the optimal set of counterfactual explanations is NP-hard. However, we further show that the corresponding objective is nondecreasing and satisfies submodularity. Therefore, a standard greedy algorithm offers an approximation factor of $(1-1/e)$ at solving the problem. Additionally, we also show that the problem of jointly finding both the optimal policy and set of counterfactual explanations reduces to maximizing a non-monotone submodular function. As a result, we can use a recent randomized algorithm to solve the problem, which offers an approximation factor of $1/e$. Finally, we illustrate our theoretical findings by performing experiments on synthetic and real lending data.