arXiv subjects

Teddy Lazebnik

Publications and source records attributed to Teddy Lazebnik.

At least 37 records · Page 2Linked to original sources

arXiv2025

Introducing 'Inside' Out of Distribution

Detecting and understanding out-of-distribution (OOD) samples is crucial in machine learning (ML) to ensure reliable model performance. Current OOD studies primarily focus on extrapolatory (outside) OOD, neglecting potential cases of interpolatory (inside) OOD. In this study, we introduce a novel perspective on OOD by suggesting it can be divided into inside and outside cases. We examine the inside-outside OOD profiles of datasets and their impact on ML model performance, using normalized Root Mean Squared Error (RMSE) and F1 score as the performance metrics on syntetically-generated datasets with both inside and outside OOD. Our analysis demonstrates that different inside-outside OOD profiles lead to unique effects on ML model performance, with outside OOD generally causing greater performance degradation, on average. These findings highlight the importance of distinguishing between inside and outside OOD for developing effective counter-OOD methods.

cs.LG↗

arXiv2025

Investigating Tax Evasion Emergence Using Dual Large Language Model and Deep Reinforcement Learning Powered Agent-based Simulation

Tax evasion, usually the largest component of an informal economy, is a persistent challenge over history with significant socio-economic implications. Many socio-economic studies investigate its dynamics, including influencing factors, the role and influence of taxation policies, and the prediction of the tax evasion volume over time. These studies assumed such behavior is given, as observed in the real world, neglecting the "big bang" of such activity in a population. To this end, computational economy studies adopted developments in computer simulations, in general, and recent innovations in artificial intelligence (AI), in particular, to simulate and study informal economy appearance in various socio-economic settings. This study presents a novel computational framework to examine the dynamics of tax evasion and the emergence of informal economic activity. Employing an agent-based simulation powered by Large Language Models and Deep Reinforcement Learning, the framework is uniquely designed to allow informal economic behaviors to emerge organically, without presupposing their existence or explicitly signaling agents about the possibility of evasion. This provides a rigorous approach for exploring the socio-economic determinants of compliance behavior. The experimental design, comprising model validation and exploratory phases, demonstrates the framework's robustness in replicating theoretical economic behaviors. Findings indicate that individual personality traits, external narratives, enforcement probabilities, and the perceived efficiency of public goods provision significantly influence both the timing and extent of informal economic activity. The results underscore that efficient public goods provision and robust enforcement mechanisms are complementary; neither alone is sufficient to curtail informal activity effectively.

cs.IR↗

arXiv2025

Tighten The Lasso: A Convex Hull Volume-based Anomaly Detection Method

Detecting out-of-distribution (OOD) data is a critical task for maintaining model reliability and robustness. In this study, we propose a novel anomaly detection algorithm that leverages the convex hull (CH) property of a dataset by exploiting the observation that OOD samples marginally increase the CH's volume compared to in-distribution samples. Thus, we establish a decision boundary between OOD and in-distribution data by iteratively computing the CH's volume as samples are removed, stopping when such removal does not significantly alter the CH's volume. The proposed algorithm is evaluated against seven widely used anomaly detection methods across ten datasets, demonstrating performance comparable to state-of-the-art (SOTA) techniques. Furthermore, we introduce a computationally efficient criterion for identifying datasets where the proposed method outperforms existing SOTA approaches.

cs.LG↗

arXiv2025

Inducing State Anxiety in LLM Agents Reproduces Human-Like Biases in Consumer Decision-Making

Large language models (LLMs) are rapidly evolving from text generators to autonomous agents, raising urgent questions about their reliability in real-world contexts. Stress and anxiety are well known to bias human decision-making, particularly in consumer choices. Here, we tested whether LLM agents exhibit analogous vulnerabilities. Three advanced models (ChatGPT-5, Gemini 2.5, Claude 3.5-Sonnet) performed a grocery shopping task under budget constraints (24, 54, 108 USD), before and after exposure to anxiety-inducing traumatic narratives. Across 2,250 runs, traumatic prompts consistently reduced the nutritional quality of shopping baskets (Change in Basket Health Scores of -0.081 to -0.126; all pFDR<0.001; Cohens d=-1.07 to -2.05), robust across models and budgets. These results show that psychological context can systematically alter not only what LLMs generate but also the actions they perform. By reproducing human-like emotional biases in consumer behavior, LLM agents reveal a new class of vulnerabilities with implications for digital health, consumer safety, and ethical AI deployment.

cs.HC↗

arXiv2025

Knowledge-integrated AutoEncoder Model

Data encoding is a common and central operation in most data analysis tasks. The performance of other models downstream in the computational process highly depends on the quality of data encoding. One of the most powerful ways to encode data is using the neural network AutoEncoder (AE) architecture. However, the developers of AE cannot easily influence the produced embedding space, as it is usually treated as a black box technique. This means the embedding space is uncontrollable and does not necessarily possess the properties desired for downstream tasks. This paper introduces a novel approach for developing AE models that can integrate external knowledge sources into the learning process, possibly leading to more accurate results. The proposed Knowledge-integrated AutoEncoder (KiAE) model can leverage domain-specific information to make sure the desired distance and neighborhood properties between samples are preservative in the embedding space. The proposed model is evaluated on three large-scale datasets from three scientific fields and is compared to nine existing encoding models. The results demonstrate that the KiAE model effectively captures the underlying structures and relationships between the input data and external knowledge, meaning it generates a more useful representation. This leads to outperforming the rest of the models in terms of reconstruction accuracy.

cs.LG↗

arXiv2025

Publishing Instincts: An Exploration-Exploitation Framework for Studying Academic Publishing Behavior and "Home Venues"

Scholarly communication is vital to scientific advancement, enabling the exchange of ideas and knowledge. When selecting publication venues, scholars consider various factors, such as journal relevance, reputation, outreach, and editorial standards and practices. However, some of these factors are inconspicuous or inconsistent across venues and individual publications. This study proposes that scholars' decision-making process can be conceptualized and explored through the biologically inspired exploration-exploitation (EE) framework, which posits that scholars balance between familiar and under-explored publication venues. Building on the EE framework, we introduce a grounded definition for "Home Venues" (HVs) - an informal concept used to describe the set of venues where a scholar consistently publishes - and investigate their emergence and key characteristics. Our analysis reveals that the publication patterns of roughly three-quarters of computer science scholars align with the expectations of the EE framework. For these scholars, HVs typically emerge and stabilize after approximately 15-20 publications. Additionally, scholars with higher h-indexes, greater number of publications, or higher academic age tend to have higher-ranking journals as their HVs.

cs.DL↗

arXiv2025

Multi-Species Prey-Predator Dynamics During a Multi-Strain Pandemic

Small and large scale pandemics are a natural phenomenon repeatably appearing throughout history, causing ecological and biological shifts in ecosystems and a wide range of their habitats. These pandemics usually start with a single strain but shortly become multi-strain due to a mutation process of the pathogen causing the epidemic. In this study, we propose a novel eco-epidemiological model that captures multi-species prey-predator dynamics with a multi-strain pandemic. The proposed model extends and combines the Lotka-Volterra prey-predator model and the Susceptible-Infectious-Recovered epidemiological model. We investigate the ecosystem's sensitivity and stability during such a multi-strain pandemic through extensive simulation relying on both synthetic cases as well as two real-world configurations. Our results are aligned with known ecological and epidemiological findings, thus supporting the adequacy of the proposed model in realistically capturing the complex eco-epidemiological properties of the multi-species multi-strain pandemic dynamics.

q-bio.PE↗

arXiv2025

Detecting LLM-assisted writing in scientific communication: Are we there yet?

Large Language Models (LLMs), exemplified by ChatGPT, have significantly reshaped text generation, particularly in the realm of writing assistance. While ethical considerations underscore the importance of transparently acknowledging LLM use, especially in scientific communication, genuine acknowledgment remains infrequent. A potential avenue to encourage accurate acknowledging of LLM-assisted writing involves employing automated detectors. Our evaluation of four cutting-edge LLM-generated text detectors reveals their suboptimal performance compared to a simple ad-hoc detector designed to identify abrupt writing style changes around the time of LLM proliferation. We contend that the development of specialized detectors exclusively dedicated to LLM-assisted writing detection is necessary. Such detectors could play a crucial role in fostering more authentic recognition of LLM involvement in scientific communication, addressing the current challenges in acknowledgment practices.

cs.IR↗

arXiv2025

Interpretable Transformation and Analysis of Timelines through Learning via Surprisability

The analysis of high-dimensional timeline data and the identification of outliers and anomalies is critical across diverse domains, including sensor readings, biological and medical data, historical records, and global statistics. However, conventional analysis techniques often struggle with challenges such as high dimensionality, complex distributions, and sparsity. These limitations hinder the ability to extract meaningful insights from complex temporal datasets, making it difficult to identify trending features, outliers, and anomalies effectively. Inspired by surprisability -- a cognitive science concept describing how humans instinctively focus on unexpected deviations - we propose Learning via Surprisability (LvS), a novel approach for transforming high-dimensional timeline data. LvS quantifies and prioritizes anomalies in time-series data by formalizing deviations from expected behavior. LvS bridges cognitive theories of attention with computational methods, enabling the detection of anomalies and shifts in a way that preserves critical context, offering a new lens for interpreting complex datasets. We demonstrate the usefulness of LvS on three high-dimensional timeline use cases: a time series of sensor data, a global dataset of mortality causes over multiple years, and a textual corpus containing over two centuries of State of the Union Addresses by U.S. presidents. Our results show that the LvS transformation enables efficient and interpretable identification of outliers, anomalies, and the most variable features along the timeline.

stat.ME↗

arXiv2025

The Academic Midas Touch: An Indicator of Academic Excellence

The recognition of academic excellence is fundamental to the scientific and academic endeavor. However, the term "academic excellence" is often interpreted in different ways, typically, using popular scientometrics such as the H-index, i10-index, and citation counts. In this work, we study an under-explored aspect of academic excellence -- researchers' propensity to produce highly cited publications. We formulate this novel perspective using a simple yet effective indicator termed the "Academic Midas Touch" (AMT). We empirically show that this perspective does not fully align with popular scientometrics and favorably compares to them in distinguishing award-winning scientists.

cs.DL↗

arXiv2025

An Empirically-parametrized Spatio-Temporal Extended-SIR Model for Combined Dilution and Vaccination Mitigation for Rabies Outbreaks in Wild Jackals

The transmission of zoonotic diseases between animals and humans poses an increasing threat. Rabies is a prominent example with various instances globally, facilitated by a surplus of meso-predators (commonly, facultative synanthropic species e.g., golden jackals [Canis aureus, hereafter jackals]) thanks to the abundance of anthropogenic resources leading to dense populations close to human establishments. To mitigate rabies outbreaks and prevent human infections, authorities target the jackal which is the main rabies vector in many regions, through the dissemination of oral vaccines in known jackals' activity centers, as well as opportunistic culling to reduce population density. Because dilution (i.e., culling) is not selective towards sick or un-vaccinated individuals, these two complementary epizootic intervention policies (EIPs) can interfere with each other. Nonetheless, there is only limited examination of the interactive effectiveness of these EIPs and their potential influence on rabies epizootic spread dynamics, highlighting the need to understand these measures and the spread of rabies in wild jackals. In this study, we introduce a novel spatio-temporal extended-SIR (susceptible-infected-recovered) model with a graph-based spatial framework for evaluating mitigation efficiency. We implement the model in a case study using a jackal population in northern Israel, and using spatial and movement data collected by Advanced Tracking and Localization of Animals in real-life Systems (ATLAS) telemetry. An agent-based simulation approach allows us to explore various biologically-realistic scenarios, and assess the impact of different EIPs configurations. Our model suggests that under biologically-realistic underlying assumptions and scenarios, the effectiveness of both EIPs is not influenced much by the jackal population size but is sensitive to their dispersal between activity centers.

cs.IR↗

arXiv2025

Data Augmentation for Deep Learning Regression Tasks by Machine Learning Models

Deep learning (DL) models have gained prominence in domains such as computer vision and natural language processing but remain underutilized for regression tasks involving tabular data. In these cases, traditional machine learning (ML) models often outperform DL models. In this study, we propose and evaluate various data augmentation (DA) techniques to improve the performance of DL models for tabular data regression tasks. We compare the performance gain of Neural Networks by different DA strategies ranging from a naive method of duplicating existing observations and adding noise to a more sophisticated DA strategy that preserves the underlying statistical relationship in the data. Our analysis demonstrates that the advanced DA method significantly improves DL model performance across multiple datasets and regression tasks, resulting in an average performance increase of over 10\% compared to baseline models without augmentation. The efficacy of these DA strategies was rigorously validated across 30 distinct datasets, with multiple iterations and evaluations using three different automated deep learning (AutoDL) frameworks: AutoKeras, H2O, and AutoGluon. This study demonstrates that by leveraging advanced DA techniques, DL models can realize their full potential in regression tasks, thereby contributing to broader adoption and enhanced performance in practical applications.

cs.LG↗

arXiv2024

Spatio-Temporal SIR Model of Pandemic Spread During Warfare with Optimal Dual-use Healthcare System Administration using Deep Reinforcement Learning

Large-scale crises, including wars and pandemics, have repeatedly shaped human history, and their simultaneous occurrence presents profound challenges to societies. Understanding the dynamics of epidemic spread during warfare is essential for developing effective containment strategies in complex conflict zones. While research has explored epidemic models in various settings, the impact of warfare on epidemic dynamics remains underexplored. In this study, we proposed a novel mathematical model that integrates the epidemiological SIR (susceptible-infected-recovered) model with the war dynamics Lanchester model to explore the dual influence of war and pandemic on a population's mortality. Moreover, we consider a dual-use military and civil healthcare system that aims to reduce the overall mortality rate which can use different administration policies. Using an agent-based simulation to generate in silico data, we trained a deep reinforcement learning model for healthcare administration policy and conducted an intensive investigation on its performance. Our results show that a pandemic during war conduces chaotic dynamics where the healthcare system should either prioritize war-injured soldiers or pandemic-infected civilians based on the immediate amount of mortality from each option, ignoring long-term objectives. Our findings highlight the importance of integrating conflict-related factors into epidemic modeling to enhance preparedness and response strategies in conflict-affected areas.

q-bio.QM↗

arXiv2024

Pulling the Carpet Below the Learner's Feet: Genetic Algorithm To Learn Ensemble Machine Learning Model During Concept Drift

Data-driven models, in general, and machine learning (ML) models, in particular, have gained popularity over recent years with an increased usage of such models across the scientific and engineering domains. When using ML models in realistic and dynamic environments, users need to often handle the challenge of concept drift (CD). In this study, we explore the application of genetic algorithms (GAs) to address the challenges posed by CD in such settings. We propose a novel two-level ensemble ML model, which combines a global ML model with a CD detector, operating as an aggregator for a population of ML pipeline models, each one with an adjusted CD detector by itself responsible for re-training its ML model. In addition, we show one can further improve the proposed model by utilizing off-the-shelf automatic ML methods. Through extensive synthetic dataset analysis, we show that the proposed model outperforms a single ML pipeline with a CD algorithm, particularly in scenarios with unknown CD characteristics. Overall, this study highlights the potential of ensemble ML and CD models obtained through a heuristic and adaptive optimization process such as the GA one to handle complex CD events.

cs.LG↗

arXiv2024

Break a Lag: Triple Exponential Moving Average for Enhanced Optimization

The performance of deep learning models is critically dependent on sophisticated optimization strategies. While existing optimizers have shown promising results, many rely on first-order Exponential Moving Average (EMA) techniques, which often limit their ability to track complex gradient trends accurately. This fact can lead to a significant lag in trend identification and suboptimal optimization, particularly in highly dynamic gradient behavior. To address this fundamental limitation, we introduce Fast Adaptive Moment Estimation (FAME), a novel optimization technique that leverages the power of Triple Exponential Moving Average. By incorporating an advanced tracking mechanism, FAME enhances responsiveness to data dynamics, mitigates trend identification lag, and optimizes learning efficiency. Our comprehensive evaluation encompasses different computer vision tasks including image classification, object detection, and semantic segmentation, integrating FAME into 30 distinct architectures ranging from lightweight CNNs to Vision Transformers. Through rigorous benchmarking against state-of-the-art optimizers, FAME demonstrates superior accuracy and robustness. Notably, it offers high scalability, delivering substantial improvements across diverse model complexities, architectures, tasks, and benchmarks.

cs.CV↗

arXiv2024

Global Lightning-Ignited Wildfires Prediction and Climate Change Projections based on Explainable Machine Learning Models

Wildfires pose a significant natural disaster risk to populations and contribute to accelerated climate change. As wildfires are also affected by climate change, extreme wildfires are becoming increasingly frequent. Although they occur less frequently globally than those sparked by human activities, lightning-ignited wildfires play a substantial role in carbon emissions and account for the majority of burned areas in certain regions. While existing computational models, especially those based on machine learning, aim to predict lightning-ignited wildfires, they are typically tailored to specific regions with unique characteristics, limiting their global applicability. In this study, we present machine learning models designed to characterize and predict lightning-ignited wildfires on a global scale. Our approach involves classifying lightning-ignited versus anthropogenic wildfires, and estimating with high accuracy the probability of lightning to ignite a fire based on a wide spectrum of factors such as meteorological conditions and vegetation. Utilizing these models, we analyze seasonal and spatial trends in lightning-ignited wildfires shedding light on the impact of climate change on this phenomenon. We analyze the influence of various features on the models using eXplainable Artificial Intelligence (XAI) frameworks. Our findings highlight significant global differences between anthropogenic and lightning-ignited wildfires. Moreover, we demonstrate that, even over a short time span of less than a decade, climate changes have steadily increased the global risk of lightning-ignited wildfires. This distinction underscores the imperative need for dedicated predictive models and fire weather indices tailored specifically to each type of wildfire.

cs.LG↗

arXiv2024

A Comprehensive Benchmark of Machine and Deep Learning Across Diverse Tabular Datasets

The analysis of tabular datasets is highly prevalent both in scientific research and real-world applications of Machine Learning (ML). Unlike many other ML tasks, Deep Learning (DL) models often do not outperform traditional methods in this area. Previous comparative benchmarks have shown that DL performance is frequently equivalent or even inferior to models such as Gradient Boosting Machines (GBMs). In this study, we introduce a comprehensive benchmark aimed at better characterizing the types of datasets where DL models excel. Although several important benchmarks for tabular datasets already exist, our contribution lies in the variety and depth of our comparison: we evaluate 111 datasets with 20 different models, including both regression and classification tasks. These datasets vary in scale and include both those with and without categorical variables. Importantly, our benchmark contains a sufficient number of datasets where DL models perform best, allowing for a thorough analysis of the conditions under which DL models excel. Building on the results of this benchmark, we train a model that predicts scenarios where DL models outperform alternative methods with 86.1% accuracy (AUC 0.78). We present insights derived from this characterization and compare these findings to previous benchmarks.

cs.LG↗

arXiv2024

The Topology of a Family Tree Graph and Its Members' Satisfaction with One Another: A Machine Learning Approach

Family members' satisfaction with one another is central to creating healthy and supportive family environments. In this work, we propose and implement a novel computational technique aimed at exploring the possible relationship between the topology of a given family tree graph and its members' satisfaction with one another. Through an extensive empirical evaluation ($N=486$ families), we show that the proposed technique brings about highly accurate results in predicting family members' satisfaction with one another based solely on the family graph's topology. Furthermore, the results indicate that our technique favorably compares to baseline regression models which rely on established features associated with family members' satisfaction with one another in prior literature.

physics.soc-ph↗