Search Results for author: Diego Klabjan

Found 79 papers, 11 papers with code

Multi-Layer Attention-Based Explainability via Transformers for Tabular Data

no code implementations28 Feb 2023 Andrea Treviño Gavito, Diego Klabjan, Jean Utke

We propose a graph-oriented attention-based explainability method for tabular data.

Gradient-Boosted Based Structured and Unstructured Learning

no code implementations28 Feb 2023 Andrea Treviño Gavito, Diego Klabjan, Jean Utke

Our proposed frameworks allow joint learning on both kinds of data by integrating the paradigms of boosting models and deep neural networks.

Second-order methods

Feature Acquisition using Monte Carlo Tree Search

no code implementations21 Dec 2022 Sungsoo Lim, Diego Klabjan, Mark Shapiro

Feature acquisition algorithms address the problem of acquiring informative features while balancing the costs of acquisition to improve the learning performances of ML models.

Pareto Regret Analyses in Multi-objective Multi-armed Bandit

no code implementations1 Dec 2022 Mengfan Xu, Diego Klabjan

We study Pareto optimality in multi-objective multi-armed bandit by providing a formulation of adversarial multi-objective multi-armed bandit and properly defining its Pareto regrets that can be generalized to stochastic settings as well.

Adversarial Attack

A Primal-Dual Algorithm for Hybrid Federated Learning

no code implementations14 Oct 2022 Tom Overman, Garrett Blum, Diego Klabjan

Furthermore, we provide experimental results that demonstrate the performance improvements of the algorithm over a commonly used method in federated learning, FedAvg.

Federated Learning

Divergence Results and Convergence of a Variance Reduced Version of ADAM

no code implementations11 Oct 2022 Ruiqi Wang, Diego Klabjan

Under a variance reduction assumption, we show that an ADAM-type algorithm converges, which means that it is the variance of gradients that causes the divergence of original ADAM.

Stochastic Optimization

Topic Analysis for Text with Side Data

no code implementations1 Mar 2022 Biyi Fang, Kripa Rajshekhar, Diego Klabjan

In the context of text, the neural network provides an overview distribution about side data for the corresponding text, which is the prior distribution in LDA to help perform topic grouping.

Tricks and Plugins to GBM on Images and Sequences

no code implementations1 Mar 2022 Biyi Fang, Jean Utke, Diego Klabjan

Convolutional neural networks (CNNs) and transformers, which are composed of multiple processing layers and blocks to learn the representations of data with multiple abstract levels, are the most successful machine learning models in recent years.

Open-Set Recognition of Breast Cancer Treatments

no code implementations9 Jan 2022 Alexander Cao, Diego Klabjan, Yuan Luo

Open-set recognition generalizes a classification task by classifying test samples as one of the known classes from training or "unknown."

Open Set Learning Robust classification

Neuron-based Pruning of Deep Neural Networks with Better Generalization using Kronecker Factored Curvature Approximation

no code implementations16 Nov 2021 Abdolghani Ebrahimi, Diego Klabjan

The proposed algorithm directs the parameters of the compressed model toward a flatter solution by exploring the spectral radius of Hessian which results in better generalization on unseen data.

Aggregation Delayed Federated Learning

1 code implementation17 Aug 2021 Ye Xue, Diego Klabjan, Yuan Luo

Federated learning is a distributed machine learning paradigm where multiple data owners (clients) collaboratively train one machine learning model while keeping data on their own devices.

BIG-bench Machine Learning Federated Learning

Logit-based Uncertainty Measure in Classification

no code implementations6 Jul 2021 Huiyu Wu, Diego Klabjan

We introduce a new, reliable, and agnostic uncertainty measure for classification tasks called logit uncertainty.


A Probabilistic Approach to Neural Network Pruning

no code implementations20 May 2021 Xin Qian, Diego Klabjan

Neural network pruning techniques reduce the number of parameters without compromising predicting ability of a network.

Network Pruning

Non-Convex Optimization with Spectral Radius Regularization

1 code implementation22 Feb 2021 Adam Sandler, Diego Klabjan, Yuan Luo

We develop a regularization method which finds flat minima during the training of deep neural networks and other machine learning models.

Classification Models for Partially Ordered Sequences

no code implementations31 Jan 2021 Stephanie Ger, Diego Klabjan, Jean Utke

Many models such as Long Short Term Memory (LSTMs), Gated Recurrent Units (GRUs) and transformers have been developed to classify time series data with the assumption that events in a sequence are ordered.

Classification General Classification +1

The Impact of the Mini-batch Size on the Dynamics of SGD: Variance and Beyond

no code implementations1 Jan 2021 Xin Qian, Diego Klabjan

We study mini-batch stochastic gradient descent (SGD) dynamics under linear regression and deep linear networks by focusing on the variance of the gradients only given the initial weights and mini-batch size, which is the first study of this nature.


Open Set Domain Adaptation by Extreme Value Theory

no code implementations22 Dec 2020 Yiming Xu, Diego Klabjan

In this paper, we tackle the open set domain adaptation problem under the assumption that the source and the target label spaces only partially overlap, and the task becomes when the unknown classes exist, how to detect the target unknown classes and avoid aligning them with the source domain.

Domain Adaptation

Concept Drift and Covariate Shift Detection Ensemble with Lagged Labels

no code implementations8 Dec 2020 Yiming Xu, Diego Klabjan

Extensive experiments on structured and unstructured data for different type of data changes establish that our method consistently outperforms the state-of-the-art methods by a large margin.

Inverse Classification with Limited Budget and Maximum Number of Perturbed Samples

no code implementations29 Sep 2020 Jaehoon Koo, Diego Klabjan, Jean Utke

In this study, we propose a new framework to solve inverse classification that maximizes the number of perturbed samples subject to a per-feature-budget limits and favorable classification classes of the perturbed samples.

Classification General Classification

Regret Bounds and Reinforcement Learning Exploration of EXP-based Algorithms

no code implementations20 Sep 2020 Mengfan Xu, Diego Klabjan

We propose a new algorithm, namely EXP4. P, by modifying EXP4 and establish its upper bound of regret in both bounded and unbounded sub-Gaussian contextual bandit settings.

reinforcement-learning Reinforcement Learning (RL)

Efficient Architecture Search for Continual Learning

no code implementations7 Jun 2020 Qiang Gao, Zhipeng Luo, Diego Klabjan

To reach these goals, we propose a novel approach named as Continual Learning with Efficient Architecture Search, or CLEAS in short.

Continual Learning Neural Architecture Search +1

Open-Set Recognition with Gaussian Mixture Variational Autoencoders

no code implementations3 Jun 2020 Alexander Cao, Yuan Luo, Diego Klabjan

In inference, open-set classification is to either classify a sample into a known class from training or reject it as an unknown class.

Classification General Classification +1

Neural Network Retraining for Model Serving

no code implementations29 Apr 2020 Diego Klabjan, Xiaofeng Zhu

We address two challenges of life-long retraining: catastrophic forgetting and efficient retraining.

Multi-Armed Bandits

The Impact of the Mini-batch Size on the Variance of Gradients in Stochastic Gradient Descent

no code implementations27 Apr 2020 Xin Qian, Diego Klabjan

The mini-batch stochastic gradient descent (SGD) algorithm is widely used in training machine learning models, in particular deep learning models.


Keyword-based Topic Modeling and Keyword Selection

no code implementations22 Jan 2020 Xingyu Wang, Lida Zhang, Diego Klabjan

The challenge is that these need to be specified ahead of knowing the forthcoming documents and the underlying topics.

Listwise Learning to Rank by Exploring Unique Ratings

1 code implementation7 Jan 2020 Xiaofeng Zhu, Diego Klabjan

We encode all of the documents already selected by an RNN model.


Conditional Hierarchical Bayesian Tucker Decomposition for Genetic Data Analysis

1 code implementation27 Nov 2019 Adam Sandler, Diego Klabjan, Yuan Luo

We apply these models to examine patients with one of four common types of cancer (breast, lung, prostate, and colorectal) and siblings with and without autism spectrum disorder.

Tensor Decomposition

Layer Flexible Adaptive Computation Time for Recurrent Neural Networks

no code implementations25 Sep 2019 Lida Zhang, Diego Klabjan

Deep recurrent neural networks perform well on sequence data and are the model of choice.

Language Modelling

Mixture-based Multiple Imputation Model for Clinical Data with a Temporal Dimension

1 code implementation12 Aug 2019 Ye Xue, Diego Klabjan, Yuan Luo

The problem of missing values in multivariable time series is a key challenge in many applications such as clinical data mining.

Gaussian Processes Imputation +1

Data Extraction from Charts via Single Deep Neural Network

no code implementations6 Jun 2019 Xiaoyi Liu, Diego Klabjan, Patrick NBless

Automatic data extraction from charts is challenging for two reasons: there exist many relations among objects in a chart, which is not a common consideration in general computer vision problems; and different types of charts may not be processed by the same model.

object-detection Object Detection

Dynamic Cell Structure via Recursive-Recurrent Neural Networks

no code implementations25 May 2019 Xin Qian, Matthew Kennedy, Diego Klabjan

In a recurrent setting, conventional approaches to neural architecture search find and fix a general model for all data samples and time steps.

Language Modelling Neural Architecture Search +1

Scale Invariant Power Iteration

no code implementations23 May 2019 Cheolmin Kim, Youngseok Kim, Diego Klabjan

In this work, we introduce a new class of optimization problems called scale invariant problems and prove that they can be efficiently solved by scale invariant power iteration (SCI-PI) with a generalized convergence guarantee of power iteration.

Convergence Analyses of Online ADAM Algorithm in Convex Setting and Two-Layer ReLU Neural Network

no code implementations22 May 2019 Biyi Fang, Diego Klabjan

We cover the convex setting showing the regret of the order of the square root of the size of the window in the constant and dynamic learning rate scenarios.

Automatic Ontology Learning from Domain-Specific Short Unstructured Text Data

no code implementations7 Mar 2019 Yiming Xu, Dnyanesh Rajpathak, Ian Gibbs, Diego Klabjan

Ontology learning is non-trivial due to several reasons with limited amount of prior research work that automatically learns a domain specific ontology from data.

General Classification Information Retrieval +1

Autoencoders and Generative Adversarial Networks for Imbalanced Sequence Classification

1 code implementation8 Jan 2019 Stephanie Ger, Yegna Subramanian Jambunath, Diego Klabjan

In addition to the medical device dataset, we also evaluate the GAN-AE performance on two additional datasets and demonstrate the application of GAN-AE to a sequence-to-sequence task where both synthetic sequence inputs and sequence outputs must be generated.

Anomaly Detection Classification +2

Layer Flexible Adaptive Computational Time

no code implementations6 Dec 2018 Lida Zhang, Abdolghani Ebrahimi, Diego Klabjan

Deep recurrent neural networks perform well on sequence data and are the model of choice.

Language Modelling

Unified recurrent network for many feature types

no code implementations27 Sep 2018 Alexander Stec, Diego Klabjan, Jean Utke

We also include two types of static (whole sequence level) features, one related to time and one not, which are combined with the encoder output.

Time Series Analysis

Unified recurrent neural network for many feature types

no code implementations24 Sep 2018 Alexander Stec, Diego Klabjan, Jean Utke

We also include two types of static (whole sequence level) features, one related to time and one not, which are combined with the encoder output.

Time Series Analysis

Nested multi-instance classification

no code implementations30 Aug 2018 Alexander Stec, Diego Klabjan, Jean Utke

We also introduce a method to replace instances that are missing which successfully creates neutral input instances and consistently outperforms standard fill-in methods in real world use cases.

Classification General Classification

Dynamic Prediction Length for Time Series with Sequence to Sequence Networks

no code implementations2 Jul 2018 Mark Harmon, Diego Klabjan

Recurrent neural networks and sequence to sequence models require a predetermined length for prediction output length.

Time Series Analysis

Forecasting Crime with Deep Learning

no code implementations5 Jun 2018 Alexander Stec, Diego Klabjan

With our best model we are able to predict the correct bin for overall crime count with 75. 6% and 65. 3% accuracy for Chicago and Portland, respectively.

Crime Prediction

Bayesian active learning for choice models with deep Gaussian processes

no code implementations4 May 2018 Jie Yang, Diego Klabjan

In this paper, we propose an active learning algorithm and models which can gradually learn individual's preference through pairwise comparisons.

Active Learning Gaussian Processes

k-Nearest Neighbors by Means of Sequence to Sequence Deep Neural Networks and Memory Networks

no code implementations27 Apr 2018 Yiming Xu, Diego Klabjan

In this paper, we propose two families of models built on a sequence to sequence model and a memory network model to mimic the k-Nearest Neighbors model, which generate a sequence of labels, a sequence of out-of-sample feature vectors and a final label for classification, and thus they could also function as oversamplers.

General Classification

Improved Classification Based on Deep Belief Networks

no code implementations25 Apr 2018 Jaehoon Koo, Diego Klabjan

For better classification generative models are used to initialize the model and model features before training a classifier.

Classification General Classification

A Stochastic Large-scale Machine Learning Algorithm for Distributed Features and Observations

no code implementations29 Mar 2018 Biyi Fang, Diego Klabjan

As the size of modern data sets exceeds the disk and memory capacities of a single computer, machine learning practitioners have resorted to parallel and distributed computing.

BIG-bench Machine Learning Distributed Computing +1

Truth Validation with Evidence

no code implementations15 Feb 2018 Papis Wongchaisuwat, Diego Klabjan

The proposed system provides supporting evidence when the statement is tagged as false.

Competitive Multi-agent Inverse Reinforcement Learning with Sub-optimal Demonstrations

1 code implementation ICML 2018 Xingyu Wang, Diego Klabjan

Compared to previous works that decouple agents in the game by assuming optimality in expert strategies, we introduce a new objective function that directly pits experts against Nash Equilibrium strategies, and we design an algorithm to solve for the reward function in the context of inverse reinforcement learning with deep neural networks as model approximations.

reinforcement-learning Reinforcement Learning (RL)

Generative Adversarial Nets for Multiple Text Corpora

1 code implementation25 Dec 2017 Baiyang Wang, Diego Klabjan

Generative adversarial nets (GANs) have been successfully applied to the artificial generation of image data.

Cross-corpus Word Embeddings

OSTSC: Over Sampling for Time Series Classification in R

no code implementations27 Nov 2017 Matthew Dixon, Diego Klabjan, Lan Wei

The OSTSC package is a powerful oversampling approach for classifying univariant, but multinomial time series data in R. This article provides a brief overview of the oversampling methodology implemented by the package.

Classification General Classification +2

A Simple and Fast Algorithm for L1-norm Kernel PCA

no code implementations28 Sep 2017 Cheolmin Kim, Diego Klabjan

We present an algorithm for L1-norm kernel PCA and provide a convergence analysis for it.

Outlier Detection

Semantic Document Distance Measures and Unsupervised Document Revision Detection

1 code implementation IJCNLP 2017 Xiaofeng Zhu, Diego Klabjan, Patrick Bless

In this paper, we model the document revision detection problem as a minimum cost branching problem that relies on computing document distances.

Dynamic Time Warping

Unsupervised Terminological Ontology Learning based on Hierarchical Topic Modeling

1 code implementation29 Aug 2017 Xiaofeng Zhu, Diego Klabjan, Patrick Bless

In this paper, we present hierarchical relationbased latent Dirichlet allocation (hrLDA), a data-driven hierarchical topic model for extracting terminological ontologies from a large number of heterogeneous documents.

Topic Models

Online Adaptive Machine Learning Based Algorithm for Implied Volatility Surface Modeling

no code implementations6 Jun 2017 Yaxiong Zeng, Diego Klabjan

In this work, we design a machine learning based method, online adaptive primal support vector regression (SVR), to model the implied volatility surface (IVS).

BIG-bench Machine Learning regression

Improving the Expected Improvement Algorithm

no code implementations NeurIPS 2017 Chao Qin, Diego Klabjan, Daniel Russo

To overcome this shortcoming, we introduce a simple modification of the expected improvement algorithm.

Diminishing Batch Normalization

no code implementations ICLR 2019 Yintai Ma, Diego Klabjan

Our proposed DBN algorithm remains the overall structure of the original BN algorithm while introduces a weighted averaging update to some trainable parameters.

Activation Ensembles for Deep Neural Networks

no code implementations24 Feb 2017 Mark Harmon, Diego Klabjan

We propose a new methodology of designing activation functions within a neural network at each layer.

An Attention-Based Deep Net for Learning to Rank

no code implementations20 Feb 2017 Baiyang Wang, Diego Klabjan

In information retrieval, learning to rank constructs a machine-based ranking model which given a query, sorts the search results by their degree of relevance or importance to the query.

Image Retrieval Information Retrieval +2

Semi-supervised Learning for Discrete Choice Models

no code implementations16 Feb 2017 Jie Yang, Sergey Shebalov, Diego Klabjan

Two classic semi-supervised learning algorithms, the expectation maximization algorithm and the cluster-and-label algorithm, have been adapted to our choice modeling problem setting.

Subset Selection for Multiple Linear Regression via Optimization

no code implementations27 Jan 2017 Young Woong Park, Diego Klabjan

For high dimensional cases, an iterative heuristic algorithm is proposed based on the mathematical programming models and a core set concept, and a randomized version of the algorithm is derived to guarantee convergence to the global optimum.


Bayesian Network Learning via Topological Order

no code implementations20 Jan 2017 Young Woong Park, Diego Klabjan

We propose a mixed integer programming (MIP) model and iterative algorithms based on topological orders to solve optimization problems with acyclic constraints on a directed graph.


Optimization for Large-Scale Machine Learning with Distributed Features and Observations

1 code implementation31 Oct 2016 Alexandros Nathan, Diego Klabjan

As the size of modern data sets exceeds the disk and memory capacities of a single computer, machine learning practitioners have resorted to parallel and distributed computing.

BIG-bench Machine Learning Distributed Computing +1

Predicting Shot Making in Basketball Learnt from Adversarial Multiagent Trajectories

no code implementations15 Sep 2016 Mark Harmon, Abdolghani Ebrahimi, Patrick Lucey, Diego Klabjan

In this paper, we predict the likelihood of a player making a shot in basketball from multiagent trajectories.

Iteratively Reweighted Least Squares Algorithms for L1-Norm Principal Component Analysis

no code implementations10 Sep 2016 Young Woong Park, Diego Klabjan

Principal component analysis (PCA) is often used to reduce the dimension of data by selecting a few orthonormal vectors that explain most of the variance structure of the data.

Regularization for Unsupervised Deep Neural Nets

no code implementations15 Aug 2016 Baiyang Wang, Diego Klabjan

Unsupervised neural networks, such as restricted Boltzmann machines (RBMs) and deep belief networks (DBNs), are powerful tools for feature selection and pattern recognition tasks.

General Classification

An Aggregate and Iterative Disaggregate Algorithm with Proven Optimality in Machine Learning

no code implementations5 Jul 2016 Young Woong Park, Diego Klabjan

We propose a clustering-based iterative algorithm to solve certain optimization problems in machine learning, where we start the algorithm by aggregating the original data, solving the problem on aggregated data, and then in subsequent steps gradually disaggregate the aggregated data.

BIG-bench Machine Learning regression

Algorithms for Generalized Cluster-wise Linear Regression

no code implementations5 Jul 2016 Young Woong Park, Yan Jiang, Diego Klabjan, Loren Williams

We examine the performance of our algorithms on a stock keeping unit (SKU) clustering problem employed in forecasting halo and cannibalization effects in promotions using real-world retail data from a large supermarket chain.


Temporal Topic Analysis with Endogenous and Exogenous Processes

no code implementations4 Jul 2016 Baiyang Wang, Diego Klabjan

We consider the problem of modeling temporal textual data taking endogenous and exogenous processes into account.


Classification-based Financial Markets Prediction using Deep Neural Networks

no code implementations29 Mar 2016 Matthew Dixon, Diego Klabjan, Jin Hoon Bang

Deep neural networks (DNNs) are powerful types of artificial neural networks (ANNs) that use several hidden layers.

Algorithmic Trading Classification +1

Going Out of Business: Auction House Behavior in the Massively Multi-Player Online Game

no code implementations24 Mar 2016 Anders Drachen, Joseph Riley, Shawna Baskin, Diego Klabjan

The in-game economies of massively multi-player online games (MMOGs) are complex systems that have to be carefully designed and managed.

Skill-Based Differences in Spatio-Temporal Team Behavior in Defence of The Ancients 2

no code implementations24 Mar 2016 Anders Drachen, Matthew Yancey, John Maguire, Derrek Chu, Iris Yuhui Wang, Tobias Mahlmann, Matthias Schubert, Diego Klabjan

Results indicate that spatio-temporal behavior of MOBA teams is related to team skill, with professional teams having smaller within-team distances and conducting more zone changes than amateur teams.

Dota 2 Time Series Clustering

Clustering Time-Series Energy Data from Smart Meters

no code implementations24 Mar 2016 Alexander Lavin, Diego Klabjan

Investigations have been performed into using clustering methods in data mining time-series data from smart meters.

Time Series Analysis

Evaluating the Performance of Offensive Linemen in the NFL

no code implementations24 Mar 2016 Nikhil Byanna, Diego Klabjan

Using a variety of statistics related to an offensive lineman's performance, we develop a framework to objectively analyze the overall performance of an individual offensive lineman and determine specific linemen who are overvalued or undervalued relative to their salary.

Semantic Properties of Customer Sentiment in Tweets

no code implementations24 Mar 2016 Eun Hee Ko, Diego Klabjan

In addition to major findings, we apply LDA (Latent Dirichlet Allocations) to the same data and drew latent topics that represent consumers' positive opinions and negative opinions on social media.

Sentiment Analysis

Predictive Analytics Using Smartphone Sensors for Depressive Episodes

no code implementations24 Mar 2016 Taeheon Jeong, Diego Klabjan, Justin Starren

We first find the home location of a patient, which is then augmented with other sensor data to identify sleep patterns and select communication patterns.

Anomaly Detection

Predicting litigation likelihood and time to litigation for patents

no code implementations23 Mar 2016 Papis Wongchaisuwat, Diego Klabjan, John O. McGinnis

We develop predictive models for estimating the likelihood of litigation for patents and the expected time to litigation based on both textual and non-textual features.

Cannot find the paper you are looking for? You can Submit a new open access paper.