Research Papers

https://doi.org/10.5281/zenodo.20711000

Abstract:
Conditional value-at-risk (CVaR), or expected shortfall, leads to large-scale constrained quadratic programs in finance, energy, logistics, contrl, and learning. Recent algorithms for CVaR-constraint quadratic programs identify projection onto a top-k-sum sublevel set as the central computational primitive. This paper studies the complementary problem: how to differentiate this projection, and how to use it as a differentiable layer.

We show that the top-k-sum projection is piecewise affine and that an active-face certificate from the forward projection determines its backward pass. On each fixed face, the Jacobian is the orthogonal projector onto the corresponding tangent space, giving exact vector-Jacobian products in O(m) time once the certificate is available. The method handles degenerate tied plateaus and provides adjoints with respect to both the projection input and the CVaR risk budget, without forming a deterministic-equivalent epigraph quadratic program. For full CVaR-constrained quadratic programs, accurate solves are differentiated through a reduced active-face KKT system, while early-stopped solves are differentiated by unrolling.

Numerical tests validate the adjoints against finite differences and an independent solver, including degenerate cases. Batched CPU/GPU implementations scale to 200 million scenarios, differentiating a 200-million-scenario projection in under 0.12 seconds on a single GPU, while deterministic-equivalent differentiation becomes impractical near 100,000 scenarios. Pre-specified reproducible experiments show that this scale is decision-relevant: hard CVaR constraints place realized out-of-sample tail risk on budget as the scenario count grows, including under distribution shift, whereas calibrated penalty methods drift or retain substantial per-instance dispersion. Code and a one-command reproduction are provided.

Keywords:
conditional value-at-risk (CVaR); top-k-sum projection; differentiable optimization; vector-Jacobian product; ADMM; risk-constrained quadratic programming

https://doi.org/10.5281/zenodo.19947043

In this monograph, I carried the central insight of my earlier paper (“Beyond Words: Mindfulness in the Diamond Sutra”) out of contemplative inquiry and into analytic philosophical argument.
(This, perhaps, is not a move that the writers of the Diamond Sutra would have made — but I wanted to test whether analytic philosophy can carry the result. This book is that test.)

(The essay can be viewed or downloaded below.)

Abstract
Some research literature have suggested that stories can influence the self-concept. We want to test whether a narrative does influence people’s notions of self in their memory or imagination. Our hypothesis is that being exposed to a story makes one reconstructs one’s self in one’s memory or imagination to be more aligned with the story’s main character. With an experiment, we exposed participants (n=129) to two conditions (control: report of facts; treatment: the facts in a story form), and then asked the Likert-scale questions to see whether they report selves in their memory or imagination that resemble the story’s main character. Using independent sample t-tests, we found no statistically significant results. Therefore, we failed to reject our Null hypothesis — there was not enough evidence to reject it.
Keywords: story, narrative, self-concept, mental time travel, memory reconstruction

Abstract
This is a planned study, despite its language. It was part of a class. In this proposed study, we examine the relation between vividness of one’s visual imagery and one’s recollection of autobiographical memories. Despite the thousands of citations of David Marks’ 1973 foundational paper of assessing one’s visual imagery vividness, only a few had looked directly at visual imagery in connection with autobiographical memories. This study seeks to fill that gap. We are the first to utilize both the Vividness of Visual Imagery Questionnaire and the Autobiographical Recollection Test in a correlational study. We studied 100 college students (49 male, 48 female, 3 non-binary or prefer-not-to-say). We found that scores of VVIQ of Marks correlates to high scores of ART. Moreover, the correlation extends to all aspects of Autobiographical Recollection as modeled by Berntsen et al. T-tests confirm the findings by showing that the effect size is noteworthy: high scorers in VVIQ have higher means of ART scores — and all 7 of ART subscales — than low scorers of VVIQ. These findings suggest, inter alia, further possible explorations into the directionality of the relation of the two constructs, into  connection between the processes underlying the generation of vivid imageries and the recollection of autobiographical memories, and into the development of clinical treatments to improve wellbeing.
Keywords: visual imagery, autobiographical memory, vividness, memory recollection, narrative coherence, narrative relevance

Abstract
Inspired by prior groundbreaking studies on maximizing by Simon and other researchers, we sought to see whether the findings extend to lower-stakes decision-making. The particular lower-stakes context we chose to study is class selection by college students. We examined the relationships between maximizing tendency and four variables: information-seeking behavior, efforting behavior, regret affect, satisfaction affect. We had 130 different participants (64% female, 32% male, 4% other) from three semesters of a particular Pscyhology class at a university. The participants took an online questionnaire, which measured their maximizing tendency, behaviors, and affects related directly to our investigation. Through correlation tests, we found that maximizing was positively correlated with information-seeking (r = .34, p < .001), with efforting (r = .37, p < .001), and with regret (r = .23, p = .01); maximizing was negatively correlated on the outcome satisfaction variable (r = -.19, p = .03). With a Median Split and independent samples t-tests, we found that maximizers score higher in every area except on outcome satisfaction, where we did not find any result statistically significant. Our findings suggest there is a relationship between maximizing and the aforesaid behaviors and affects.
Keywords: maximizing, satisficing