arXiv subjects

Teddy Lazebnik

Publications and source records attributed to Teddy Lazebnik.

58 records · Page 4Linked to original sources

arXiv2022

High Resolution Spatio-Temporal Model for Room-Level Airborne Pandemic Spread

Airborne pandemics have caused millions of deaths worldwide, large-scale economic losses, and catastrophic sociological shifts in human history. Researchers have developed multiple mathematical models and computational frameworks to investigate and predict the pandemic spread on various levels and scales such as countries, cities, large social events, and even buildings. However, modeling attempts of airborne pandemic dynamics on the smallest scale, a single room, have been mostly neglected. As time indoors increases due to global urbanization processes, more infections occur in shared rooms. In this study, a high-resolution spatio-temporal epidemiological model with airflow dynamics to evaluate airborne pandemic spread is proposed. The model is implemented using high-resolution 3D data obtained using a light detection and ranging (LiDAR) device and computing the model based on the Computational Fluid Dynamics (CFD) model for the airflow and the Susceptible-Exposed-Infected (SEI) model for the epidemiological dynamics. The pandemic spread is evaluated in four types of rooms, showing significant differences even for a short exposure duration. We show that the room's topology and individual distribution in the room define the ability of air ventilation to reduce pandemic spread throughout breathing zone infection.

cs.IR↗

arXiv2022

Academic Co-authorship is a Risky Game

Conducting a research project with multiple participants is a complex task that involves not only scientific but also multiple social, political, and psychological interactions. This complexity becomes particularly evident when it comes to navigating the selection process for the number and order of co-authors on the resulting manuscript for publication due to the current form of collaboration dynamics common in academia. There is currently no computational model to generate a data-driven suggestion that could be used as a baseline for understating these dynamics. To address this limitation, we have developed a first game-theory-based model to generate such a baseline for co-authorship. In our model, co-authors can issued an ultimatum to pause the publication of the manuscript until the underlying issue has been resolved. We modeled the effect of issuing one or more ultimatums and showed that they have a major impact on the ultimate number and position of co-authors and the length of the publication process. In addition, we explored the effect of two common relationships (student-advisor and colleague-colleague) on co-authorship scenarios. The results of our model are alarming and suggest that the current academic practices are not fit for purpose. Where they work, they work because of the integrity of researchers and not by a systematic design.

cs.DL↗

arXiv2022

Rivendell: Project-Based Academic Search Engine

Finding relevant research literature in online databases is a familiar challenge to all researchers. General search approaches trying to tackle this challenge fall into two groups: one-time search and life-time search. We observe that both approaches ignore unique attributes of the research domain and are affected by concept drift. We posit that in searching for research papers, a combination of a life-time search engine with an explicitly-provided context (project) provides a solution to the concept drift problem. We developed and deployed a project-based meta-search engine for research papers called Rivendell. Using Rivendell, we conducted experiments with 199 subjects, comparing project-based search performance to one-time and life-time search engines, revealing an improvement of up to 12.8 percent in project-based search compared to life-time search.

cs.IR↗

arXiv2022

SubStrat: A Subset-Based Strategy for Faster AutoML

Automated machine learning (AutoML) frameworks have become important tools in the data scientists' arsenal, as they dramatically reduce the manual work devoted to the construction of ML pipelines. Such frameworks intelligently search among millions of possible ML pipelines - typically containing feature engineering, model selection and hyper parameters tuning steps - and finally output an optimal pipeline in terms of predictive accuracy. However, when the dataset is large, each individual configuration takes longer to execute, therefore the overall AutoML running times become increasingly high. To this end, we present SubStrat, an AutoML optimization strategy that tackles the data size, rather than configuration space. It wraps existing AutoML tools, and instead of executing them directly on the entire dataset, SubStrat uses a genetic-based algorithm to find a small yet representative data subset which preserves a particular characteristic of the full data. It then employs the AutoML tool on the small subset, and finally, it refines the resulted pipeline by executing a restricted, much shorter, AutoML process on the large dataset. Our experimental results, performed on two popular AutoML frameworks, Auto-Sklearn and TPOT, show that SubStrat reduces their running times by 79% (on average), with less than 2% average loss in the accuracy of the resulted ML pipeline.

cs.LG↗