A locally-built, LLM-digested index of recent arXiv papers in quant finance, geometry/topology, and statistical ML — keyword search served straight from SQLite on this machine.
Behavioral cloning fails due to ignoring causal structure, leading to worse performance.
We show that stochastic recovery always leads to counter-intuitive behaviors in the risk measures of a CDO tranche - namely, continuity on default and positive credit spread risk cannot be ensured simultaneously. We then propose a simple recovery variance regularization method to control the magnitude of negative credi…
Fairness in ML models can lead to counterintuitive predictions.
Study of multi-task semi-supervised learning in high dimensions.
The paper analyzes PLS-SVD in high-dimensional data integration, revealing its strengths and limitations.
Generative adversarial training can be generally understood as minimizing certain moment matching loss defined by a set of discriminator functions, typically neural networks. The discriminator set should be large enough to be able to uniquely identify the true distribution (discriminative), and also be small enough to …
Unified model predicts structure of neural network loss landscapes.
Study analyzes EU ETS carbon market dynamics, revealing inefficiencies and anomalies.
Neural networks are vulnerable to small adversarial perturbations. Existing literature largely focused on understanding and mitigating the vulnerability of learned models. In this paper, we demonstrate an intriguing phenomenon about the most popular robust training method in the literature, adversarial training: Advers…
We present an elementary analysis of the dynamical aspects of the GDP / government surplus multiplier with relevance to the assessment of a country's debt repayment policy. We show the (at first) counter intuitive result that in order to reduce the Debt/GDP ratio, countries with high Debt to GDP should go into further …
Fuelled by increasing computer power and algorithmic advances, machine learning techniques have become powerful tools for finding patterns in data. Since quantum systems produce counter-intuitive patterns believed not to be efficiently produced by classical systems, it is reasonable to postulate that quantum computers …
The motivations for using variational inference (VI) in neural networks differ significantly from those in latent variable models. This has a counter-intuitive consequence; more expressive variational approximations can provide significantly worse predictions as compared to those with less expressive families. In this …
Importance sampling is widely used in machine learning and statistics, but its power is limited by the restriction of using simple proposals for which the importance weights can be tractably calculated. We address this problem by studying black-box importance sampling methods that calculate importance weights for sampl…
BatchNorm helps train quantized networks by avoiding gradient explosion.
EXPO framework eliminates need for reward model, achieving better optimization.
We review and illustrate how the volatility smile translates into a probability distribution, the market-implied probability distribution representing believes priced in. The effects of changes in the smile are examined. Special attention is given to the effects of slope, which might appear at first counter-intuitive. …
The paper presents a method to reduce computational and storage costs in PCA and spectral clustering.
We prove that some relative character varieties of the fundamental group of a punctured sphere into the Hermitian Lie groups admit compact connected components. The representations in these components have several counter-intuitive properties. For instance, the image of any simple closed curve is an …
MIME uses mutual information minimization for better exploration in environments with abrupt transitions.
Real-valued word representations have transformed NLP applications; popular examples are word2vec and GloVe, recognized for their ability to capture linguistic regularities. In this paper, we demonstrate a {\em very simple}, and yet counter-intuitive, postprocessing technique -- eliminate the common mean vector and a f…
Multiple-step lookahead policies have demonstrated high empirical competence in Reinforcement Learning, via the use of Monte Carlo Tree Search or Model Predictive Control. In a recent work \cite{efroni2018beyond}, multiple-step greedy policies and their use in vanilla Policy Iteration algorithms were proposed and analy…
Study on heavy tails in closing auction returns, explaining imbalance through limit order submission.
New method accounts for hidden context in preference learning for RLHF models.
Computer Vision and machine learning methods were previously used to reveal screen presence of genders in TV and movies. In this work, using head pose, gender detection, and skin color estimation techniques, we demonstrate that the gender disparity in TV in a South Asian country such as Bangladesh exhibits unique chara…
We derive expressions for the predicitive information rate (PIR) for the class of autoregressive Gaussian processes AR(N), both in terms of the prediction coefficients and in terms of the power spectral density. The latter result suggests a duality between the PIR and the multi-information rate for processes with mutua…
Markowitz's celebrated mean--variance portfolio optimization theory assumes that the means and covariances of the underlying asset returns are known. In practice, they are unknown and have to be estimated from historical data. Plugging the estimates into the efficient frontier that assumes known parameters has led to p…
The paper analyzes how momentum affects convergence in stochastic gradient methods.
Neural Shadow-Mapping uncovers causal links in dynamic systems.
Sharp bounds on ERM's minimal error in regression.
We consider a binary sequence generated by thresholding a hidden continuous sequence. The hidden variables are assumed to have a compound symmetry covariance structure with a single parameter characterizing the common correlation. We study the parameter estimation problem under such one-parameter models. We demonstrate…
Avoids resentment in classifier fairness by using monotonic models.
Sideways trains video models by overwriting activations as new frames arrive, potentially improving generalization.
New model reveals significant impact of data and parameter variations on machine learning benchmarks.
Investments with best performance are not associated with best Sharpe ratios.
How can we control for latent discrimination in predictive models? How can we provably remove it? Such questions are at the heart of algorithmic fairness and its impacts on society. In this paper, we define a new operational fairness criteria, inspired by the well-understood notion of omitted variable-bias in statistic…
New findings suggest deep generative models can misclassify outliers, requiring new evaluation methods.
Study competitive agents' optimal consumption and investment strategies with relative performance criteria.
Deep GNNs and self-supervision boost graph learning at scale.
We study the fundamental tradeoffs between statistical accuracy and computational tractability in the analysis of high dimensional heterogeneous data. As examples, we study sparse Gaussian mixture model, mixture of sparse linear regressions, and sparse phase retrieval model. For these models, we exploit an oracle-based…
We introduce DQFIM to quantify and improve generalization of quantum machine learning models.
Proposes methods to make data streams fair without fixing a model.
In this study, the wind data series from five locations in Aegean Sea islands, the most active `hotspots' in terms of refugee influx during the Oct/2015 - Jan/2016 period, are investigated. The analysis of the three-per-site data series includes standard statistical analysis and parametric distributions, auto-correlati…
Complex-valued neural networks improve wireless fingerprinting robustness.
Adversarial training enhances model transferability without sacrificing accuracy.
Model shows how confidence feedback can lead to different crisis outcomes.
Gamblers lose in long bets despite casino claims, study shows.
New metric captures individual neuron tuning across neural networks.
This work challenges the assumption that shorter conformal prediction intervals are always better.