Satellite imagery improves house price prediction models.
problem Improving accuracy of housing price estimation models.
method Transfer learning from ImageNet-pretrained Inception-v3 model to satellite images.
result Achieved a 10% improvement in R-squared score.
Deep learning predicts crime patterns in real time.
problem Accurate real-time crime prediction for public safety.
method Adapted deep learning spatio-temporal predictor ST-ResNet to predict crime distribution over Los Angeles.
result Models reveal highly accurate predictive power over a half year period.
Model predicts crime distribution in real-time, superior to existing methods.
problem Accurate real-time crime forecasting is difficult due to sparse and weak historical data.
method Adapted spatial temporal residual network to predict crime distribution in Los Angeles.
result The proposed model outperforms existing approaches in crime prediction accuracy.
Bayesian method for imputing actigraph data from mobile devices.
problem Imputing missing actigraph data from mobile devices.
method Bayesian inference and hierarchical dynamic linear model.
result Statistical learning of time-varying impact of explanatory variables on acceleration.
Method estimates travel times on urban roads using Uber data.
problem Estimating travel times on urban roads where data is scarce.
method Graph representation, trip sampling, least-squares optimization.
result Estimates travel times on arterial roads using aggregated Uber data.
New method detects conflicts in police body-worn audio.
problem Poor metrics for conflict detection in police interactions.
method Adaptive noise removal, non-speech filtering, and new phrase-based measures.
result Demonstrated effectiveness on LAPD body-worn audio data.
This text is a set of lecture notes for a series of four talks given at I.P.A.M., Los Angeles, on March 18-20, 2003. The first lecture provides a quick overview of symplectic topology and its main tools: symplectic manifolds, almost-complex structures, pseudo-holomorphic curves, Gromov-Witten invariants and Floer homol…
We demonstrate a simple strategy to cope with missing data in sequential inputs, addressing the task of multilabel classification of diagnoses given clinical time series. Collected from the pediatric intensive care unit (PICU) at Children's Hospital Los Angeles, our data consists of multivariate time series of observat…
There is growing interest in applying machine learning methods to Electronic Medical Records (EMR). Across different institutions, however, EMR quality can vary widely. This work investigated the impact of this disparity on the performance of three advanced machine learning algorithms: logistic regression, multilayer p…
hood2vec identifies urban area similarity via mobility networks.
problem Identifying similar urban areas using mobility networks.
method Learning node embeddings of the mobility network from Foursquare check-ins.
result Mobility dynamics capture different aspects of urban area similarity than venue types.
Body-worn video (BWV) cameras are increasingly utilized by police departments to provide a record of police-public interactions. However, large-scale BWV deployment produces terabytes of data per week, necessitating the development of effective computational methods to identify salient changes in video. In work carried…
Study uses machine learning to optimize seismic design parameters.
problem Optimizing seismic design parameters for performance-based design.
method Implementing explainable machine learning models to map design variables and performance metrics, integrated into a genetic optimization algorithm.
result Highly accurate surrogate models (R2> 90%) across diverse building types and hazards, identifying optimal member properties.
The paper uses CPI growth rates to improve LGD predictions for CRE loans.
problem Challenges in forecasting LGD for CRE loans due to extended resolution times and restricted data.
method Combines internal and public data, including CPI growth rates, to forecast CRE LGD.
result Incorporating CPI at the time of default improves LGD prediction accuracy.
A new metric space model for point process excitations uncovers hidden interactions.
problem Estimating pairwise interactions in multivariate Hawkes processes is often infeasible.
method Developed a Hidden Hawkes Geometry (HHG) model to embed event types in a metric space.
result Learning the embedding reveals salient interactions in various applications.
New tools analyze the complexity of left-ordering equivalence relations in groups and 3-manifolds.
problem Analyzing the complexity of conjugacy equivalence relations in left-orderable groups and 3-manifolds.
method Developed new tools to analyze the complexity of the conjugacy equivalence relation Elo(G) for left-orderable groups G. Used these tools to demonstrate non-smoothness and initiate a systematic analysis of Elo(π1(M)) for 3-manifolds. result Proved that if M is not prime, then Elo(π1(M)) is a universal countable Borel equivalence relation, and showed that in certain cases the complexity of Elo(π1(M)) is bounded below by the complexity of the conjugacy equivalence relation arising from the fundamental group of each of the JSJ pieces of M. Also proved that if M is the complement of a nontrivial knot in S3, then Elo(π1(M)) is not smooth, and showed how determining smoothness of Elo(π1(M)) for all knot manifolds M is related to the L-space conjecture. Study investigates deep learning for scalable long-term traffic predictions in large transport networks.
problem Scalable long-term traffic predictions in large transport networks.
method Investigated deep learning models for link-based predictions, including clustering and graph convolutional approaches.
result Deep learning models can be useful for long-term large-scale traffic prediction, but simpler predictors are satisfactory for short-term forecasting.
A novel spatio-temporal graph neural network with a learnable Tweedie head improves vessel traffic flow prediction in sparse maritime data.
problem Accurate vessel traffic flow prediction in sparse maritime data.
method A model-agnostic learnable Tweedie head attached to ST-GNN backbones.
result The proposed head consistently improves RMSE across multiple ST-GNN backbones, especially on non-zero events.
New model improves option pricing with faster convergence and better generalization.
problem Improving classical option pricing models.
method Introducing a time value related decision function and proving a universal approximation theorem.
result The new decision function approximates on the entire domain of definition by neural networks.
Margin maximization in the hard-margin sense, proposed as feature elimination criterion by the MFE-LO method, is combined here with data radius utilization to further aim to lower generalization error, as several published bounds and bound-related formulations pertaining to lowering misclassification risk (or error) pe…
Research aims to predict fallen angel bonds' bankruptcy using machine learning.
problem Predicting which fallen angel bonds will become investment grade or go bankrupt.
method Used four classification methods (logistic regression, KNN, SVM, NN) and Google Cloud's automated machine learning.
result Google Cloud's machine learning model performed best in over-sampled and feature selection data sets.
Proves existence of minimal surfaces of arbitrary genus with two ends.
problem Proves existence of complete minimal surfaces of arbitrary genus with specific properties.
method Extends orthodisk method with partial symmetry to introduce controlled symmetry.
result Proves existence of Angel surfaces of arbitrary genus.
This paper presents practical methods for portfolio selection in investments.
problem Investment portfolio selection challenges.
method Mean-variance optimization, mean-semivariance model, genetic algorithms, transaction costs.
result More comprehensive risk and return analysis in portfolio selection.
Olympic Games remain costly and overrun, despite reforms.
problem High costs and cost overruns in Olympic Games.
method Analysis of past and current Olympic Games costs and overruns, assessment of IOC reforms, and comparison of reuse vs. new construction.
result Cost overruns are increasing, and reuse of venues is ineffective.
The so-called risk diversification principle is analyzed, showing that its convenience depends on individual characteristics of the risks involved and the dependence relationship among them. ----- Se analiza el principio de diversificación de riesgos y se demuestra que no siempre resulta mejor que no diversificar, pues…
The paper improves ALO for ℓ1-regularized models.
problem Estimating out-of-sample error for ℓ1-regularized models. method Developed a novel theory for ℓ1-regularized problems, bounding ALO error. result For ℓ1-regularized problems, ALO error goes to zero as p goes to infinity. This paper improves indoor positioning accuracy by deploying reference nodes to ensure Line-of-Sight.
problem Systematic bias errors in indoor positioning due to non-LoS propagation.
method Model indoor service area as a graph, partition into cliques for reference nodes, set minimum distance and angle parameters.
result Guaranteed LoS to reference nodes improves indoor positioning accuracy and precision.
The paper analyzes LOCV for high-dimensional risk estimation, proving error bounds.
problem Estimating out-of-sample prediction error in high-dimensional settings.
method Theoretical analysis of leave-one-out cross validation (LOCV) in penalized regression.
result Finite sample upper bounds on LOCV error, showing it converges to zero as n,p → ∞.
Paper proposes a graph model for optimal AP deployment in indoor optical wireless networks.
problem Challenges in deploying optical wireless networks due to LoS requirement and limited range.
method Graph modeling approach to identify minimum number of APs and their optimal locations.
result Optimal deployment of APs ensures connectivity and minimizes interference in indoor environments.
Proposes AtCoR for predicting bike station usage, improving station network reconfiguration.
problem Challenges in predicting new bike stations due to lack of historical data.
method AtCoR algorithm that predicts both existing and new bike stations using station-centered heatmaps and historical correlations.
result AtCoR outperforms existing models in predicting bike station usage.
We present examples, both compact and non-compact complete, of lo- cally non-homogeneous proper A-manifolds.
Predicts physiologically acceptable states for pediatric ICU discharge.
problem Determining physiologically stable states for pediatric ICU discharge.
method Computed PASS values from hr, sbp, and dbp measurements, compared to age-normal and polynomial regression predictions.
result RNN model predictions were more accurate than age-normal vitals.
Lo-Hp decouples weight generation into local and global policies to improve flexibility and efficiency.
problem Over-coupling and long-horizon issues in current optimization methods.
method Hybrid-Policy Sub-Trajectory Balance objective.
result Learning local optimization policies addresses long-horizon issues and enhances global weight generation.
Transfer learning improves highway traffic forecasting using graph neural networks.
problem Lack of historical data for traffic forecasting on large highway networks.
method Developed a transfer learning approach for DCRNN, a graph neural network for highway forecasting.
result TL-DCRNN can forecast traffic on unseen regions of the highway network with high accuracy.
A new model improves homogeneity in burn patient reimbursement.
problem Incomplete homogeneity checks for burn patients using LOS as a proxy.
method Cost-sensitive decision tree model considering patient-level cost and severity of burn.
result Identified groups with increased homogeneity compared to current HRG groups.
In this paper, we study hyperkahler metric and practice GMN's construction of hyperkahler metric on focus-focus fibrations. We explicitly compute the action-angel coordinates on the local model of focus-focus fibration, and show its semi-global invariant should be harmonic to admit a compatible holomorphic 2-form. Then…
In arXiv:1207.0332 [cs.LO] was proposed a graphic lambda calculus formalism, which has sectors corresponding to untyped lambda calculus and emergent algebras. Here we explore the sector covering knot diagrams, which are constructed as macros over the graphic lambda calculus.
We present an avatar of the Euler obstruction to foliated structures on certain non-metric surfaces. This adumbrates (at least for the simplest 2D-configurations) that the standard mechanism---to the effect that the devil of algebra sometimes barricades the existence of angelic geometric structures (obstruction theory …
We find the minimum dilatation of pseudo-Anosov braids on n-punctured discs for 3 <= n <= 8. This covers the results of Song-Ko-Los (n=4) and Ham-Song (n=5). The proof is elementary, and uses the Lefschetz formula.
The article studies random infinite ideal hyperbolic polyhedra and their dual graphs, establishing new boundary theories.
problem Uniformization and boundary theory of random infinite ideal hyperbolic polyhedra and their dual graphs.
method Combinatorics, geometry, analysis, and random walks perspectives.
result Characterization of the ICP type of IAG and convergence of simple random walk to the boundary.
Study improves understanding of non-differentiable penalties in high-dimensional settings.
problem Theoretical understanding of non-differentiable penalties like generalized LASSO and nuclear norm in high-dimensional settings.
method Proportional high-dimensional regime analysis with finite sample upper bounds on expected squared error.
result LO provides accurate estimation of out-of-sample risk in high-dimensional settings.
The present paper provides a multi-period contagion model in the credit risk field. Our model is an extension of Davis and Lo's infectious default model. We consider an economy of n firms which may default directly or may be infected by other defaulting firms (a domino effect being also possible). The spontaneous defau…
Matrix factorization generates investment recommendations for investors.
problem Generating accurate investment recommendations for investors.
method Used matrix factorization and an iterative conjugate gradient method to optimize investment recommendations.
result Achieved highest average prediction accuracy of 13.3% for investors.
This is a draft of a book submitted for publication by the AMS. Its theme is the remarkable interplay, accelerating in the last few decades, between topology and the theory of orderable groups, with applications in both directions. It begins with an introduction to orderable groups and their algebraic properties. Many …
We prove that for every compactum X and every integer n≥2 there are a compactum Z of dim≤n+1 and a surjective UVn−1-map $r: Z \lo X$ such that for every abelian group G and every integer k≥2 such that dimGX≤k≤n we have dimGZ≤k and r is G-acyclic.
Proves Gannon-Lee theorem for C1 spacetimes.
problem Classical singularity theorems for C1 spacetimes. method Proves theorem for C1 spacetimes, shows geodesic properties. result Gannon-Lee theorem holds for C1 spacetimes. We prove that the link of a complex normal surface singularity is an L--space if and only if the singularity is rational. This via a recent result of Hanselman, J. Rasmussen, S. D. Rasmussen and Watson (proving the conjecture of Boyer, Gordon and Watson), shows that a singularity link is not rational if and only if its…
Motivated by well known results in low-dimensional topology, we introduce and study a topology on the set CO(G) of all left-invariant circular orders on a fixed countable and discrete group G. CO(G) contains as a closed subspace LO(G), the space of all left-invariant linear orders of G, as first topologized by Sikora. …
The aim of this paper is to extend the notion of pseudo harmonic morphism (introduced by Loubeau \cite {Lo}) to the case when the source manifold is an admissible Riemannian polyhedron. We define these maps to be harmonic in the sense of Eells-Fuglede \cite {EF} and pseudo-horizontally weakly conformal in our sense (se…