Soo Jin (Suzie) Kim

The assumption that the justifiability of scientific belief depends exclusively on the relevant facts is a widely accepted orthodoxy both inside and outside of the scientific establishment. Drawing on the pragmatic and moral encroachment thesis in epistemology, this article challenges that assumption by showing that practical and moral considerations affect the justifiability of our beliefs regarding the safety of COVID-19 vaccines. In particular, I show that a proper consideration of the practical and moral costs of committing to the belief that COVID-19 vaccines are safe enough for a general vaccine mandate requires COVID-19 vaccines to meet a higher-than-current evidentiary standard of safety. Furthermore, I show that a reassessment of existing COVID-19 vaccines against a higher evidentiary standard of safety would entail a wider scope of medical exemptions and remedial duties. Beyond the case of COVID-19 vaccines, this argument has important general implications for how scientific beliefs should be assessed.

Discovering the underlying mathematical expressions describing a dataset is a core challenge for artificial intelligence. This is the problem of symbolic regression. Despite recent advances in training neural networks to solve complex tasks, deep learning approaches to symbolic regression are underexplored. We propose a framework that leverages deep learning for symbolic regression via a simple idea: use a large model to search the space of small models. Specifically, we use a recurrent neural network to emit a distribution over tractable mathematical expressions and employ a novel risk-seeking policy gradient to train the network to generate better-fitting expressions. Our algorithm outperforms several baseline methods (including Eureqa, the gold standard for symbolic regression) in its ability to exactly recover symbolic expressions on a series of benchmark problems, both with and without added noise. More broadly, our contributions include a framework that can be applied to optimize hierarchical, variable-length objects under a blackbox performance metric, with the ability to incorporate constraints in situ, and a risk-seeking policy gradient formulation that optimizes for best-case performance instead of expected performance.

Supporters of paternalistic policies argue that interference with risky or dangerous choices for citizens’ own good is permissible, as long as those choices are caused by cognitive irrationality or ignorance. Yet, some liberal thinkers argue that despite human irrationality, paternalistic policies are still wrong because they fail to respect citizens as moral equals. I argue that actually both views are mistaken about what respect for citizens requires, because they conceptualize the citizens’ interests from the wrong standpoint. In order for citizens to be respected as equals, I argue that citizens’ interests must be defined from a joint (second-person) standpoint which is constructed through a dialogical process between policymakers and citizens oriented towards mutual understanding. Furthermore, I argue that engaging citizens in such a dialogue is a distinctive paternalistic intervention in its own right, which unlike other kinds of paternalistic intervention, is compatible with respect for citizens as equals.

According to T.M. Scanlon, the core idea of contractualism consists in the claim that what we are morally required to do is conceptually grounded in the value of living in “relations of mutual recognition” with others. Specifically, Scanlon's contractualist idea of “living in relations of mutual recognition with others” requires that one act only in ways that cannot be reasonably rejected by all of those affected, according to the results of a hypothetical reflection conducted within one's own mind. I claim, however, that Scanlon is mistaken to rely on a single individual's hypothetical reflection (monological reflection) to determine what no one affected could reasonably reject. Instead, following Jürgen Habermas, I claim that we must engage in actual dialogue with all who stand to be affected, in order to live in relations of mutual recognition with others.

One of the most salient objections against paternalism is that it is motivated by a negative judgment about other people’s capacity to advance their own goals and interests. Such a negative judgment, according to this objection, is morally wrong because it denies others the status of moral equals who can rationally set and pursue their own conception of the good. Despite the popularity of this objection, I argue that it misfires because rendering a negative judgment about others’ capacities does not deny them equal moral status when there is sufficient _evidence_ for that judgment. However, I argue that in order to determine if there is sufficient evidence of the agent’s incapacity, we must understand the agent’s reasons for action through a rational discursive exchange with them, instead of taking perceived facts about their behavior as sufficient grounds for a negative (or positive) judgment of their capacities. In fact, I argue that if the outcome of the discursive exchange supports a negative judgment of the agent’s capacities, there is a _pro tanto_ case in favor of paternalistic interference.