arXiv research

The paper argues for using Neyman orthogonal score for balancing in debiased machine learning.

We develop an axiomatic theory of balance functions (future value functions) in the theory of interest that is derived from financial considerations and which applies to general regulated payment streams, including continuous payment streams. Balance functions exist and are unique up to an initial choice of deposit and…

Balanced Activation improves object detection performance on long-tailed datasets.

Researchers derive a new equation for valuing American options.

Cross-balancing improves causal inference by balancing features with outcome data.

Classifies Bitcoin addresses based on their balance functions.

Unified theory for causal inference using various methods.

Kernel balancing weights are generalized as KRRR, providing better confidence intervals for treatment effects.

We analyze the classical model of compound interest with a constant per-period payment and interest rate. We examine the outstanding balance function as well as the periodic payment function and show that the outstanding balance function is not generally concave in the interest rate, but instead may be initially convex…

Proposes a balanced multi-component and multi-layer neural network for efficient function approximation.

For a certain maximal unipotent family of Abelian varieties over the punctured disc, we show that after a base change, one can complete the family over a disc such that the whole degeneration can be simultaneously balanced embedded into a projective space by the theta functions. Then we study the relationship between t…

Extends Gauduchon's result to higher dimensions, showing balanced metrics.

Paper simplifies balancing weights by relaxing outcome assumptions.

The paper studies matrix normalization and graph balancing using a new functional and gradient descent.

Paper surveys balanced metrics and proves a geodesic convexity result.

LSB is a new MCMC method for discrete spaces that reduces target evaluations.

BUDS balances privacy and utility by shuffling data, achieving strong privacy with minimal loss.

In Minkowski geometry the unit ball is a compact convex body $K$ containing the origin in its interior. The boundary of the body is formed by the unit vectors. We also have a so-called Minkowski functional to measure the length of vectors. By changing the origin in the interior of the body we have a smoothly varying fa…

New stability criterion for Fano manifolds using anticanonically balanced metrics.

New method for causal inference with complex treatment compositions.

New approach finds Kähler metrics on compact complex manifolds.

Let $Φ$ be a strictly plurisubharmonic and radial function on the unit disk ${\cal D}\subset {\complex}$ and let $g$ be the \K metric associated to the \K form $ω=\frac{i}{2}\partial\bar\partialΦ$. We prove that if $g$ is $g_{eucl}$-balanced of height 3 (where $g_{eucl}$ is the standard Euclidean metric on ${\complex}=…

This paper introduces new loss functions for balanced multi-class classification.

New approach finds analytic interpretation of algebraic invariants for balanced metrics.

A new method corrects bias in causal inference by balancing covariate distributions.

Noise balance theory explains SGD's behavior in neural networks.

Neurons and networks in the cerebral cortex must operate reliably despite multiple sources of noise. To evaluate the impact of both input and output noise, we determine the robustness of single-neuron stimulus selective responses, as well as the robustness of attractor states of networks of neurons performing memory ta…

Paper introduces Balanced Meta-Softmax for better long-tailed visual recognition.

This work improves structured prediction by learning the balance between signal and random noise.

Unified formula for training dynamics of linear networks combining lazy and balanced regimes.

The paper solves a problem related to curvature in complex geometry.

Optimizes Metropolis-Hastings algorithms for efficient sampling in high dimensions.

We study the asymptotic behavior of quantized Ding functionals along Bergman geodesic rays and prove that the slope at infinity can be expressed in terms of Donaldson-Futaki invariants and Chow weights. Based on the slope formula, we introduce a new algebro-geometric stability on Fano manifolds and show that the existe…

Develops a direct debiased machine learning framework using Bregman divergence.

Compared to in-clinic balance training, in-home training is not as effective. This is, in part, due to the lack of feedback from physical therapists (PTs). Here, we analyze the feasibility of using trunk sway data and machine learning (ML) techniques to automatically evaluate balance, providing accurate assessments out…

MCNet improves uncertainty calibration in online advertising by modeling complex relations and balancing performance.

Study shows sudden loss of balance in stock market networks after 2011, reducing predictability.

Generalized Thurston's characterization for branched coverings of the 2-sphere.

SurvCaus improves survival CATE estimation using neural nets.

We test a historical price time series in a financial market (the NASDAQ 100 index) for a statistical property known as detailed balance. The presence of detailed balance would imply that the market can be modeled by a stochastic process based on a Markov chain, thus leading to equilibrium. In economic terms, a positiv…

New method stabilizes IF-based estimators for causal mediation analysis with continuous mediators.

CwA optimizes search performance by jointly learning a balanced database partition and a neural probing function.

In multi-task learning, difficulty levels of different tasks are varying. There are many works to handle this situation and we classify them into five categories, including the direct sum approach, the weighted sum approach, the maximum approach, the curriculum learning approach, and the multi-objective optimization ap…

Deep Q-Learning optimizes market making by balancing price risk and spread profits.

The paper analyzes systoles of complex projective spaces under various metrics.

sgboost reduces variable selection bias in boosting with balanced group selection.

We study Smoothed Online Convex Optimization, a version of online convex optimization where the learner incurs a penalty for changing her actions between rounds. Given a $Ω(\sqrt{d})$ lower bound on the competitive ratio of any online algorithm, where $d$ is the dimension of the action space, we ask under what conditio…

PNCs balance tractability and expressiveness in probabilistic modeling.