arXiv subjects

Eitan Farchi

Publications and source records attributed to Eitan Farchi.

48 records · Page 3Linked to original sources

arXiv2021

Detecting model drift using polynomial relations

Machine learning models serve critical functions, such as classifying loan applicants as good or bad risks. Each model is trained under the assumption that the data used in training and in the field come from the same underlying unknown distribution. Often, this assumption is broken in practice. It is desirable to identify when this occurs, to minimize the impact on model performance. We suggest a new approach to detecting change in the data distribution by identifying polynomial relations between the data features. We measure the strength of each identified relation using its R-square value. A strong polynomial relation captures a significant trait of the data which should remain stable if the data distribution does not change. We thus use a set of learned strong polynomial relations to identify drift. For a set of polynomial relations that are stronger than a given threshold, we calculate the amount of drift observed for that relation. The amount of drift is measured by calculating the Bayes Factor for the polynomial relation likelihood of the baseline data versus field data. We empirically validate the approach by simulating a range of changes, and identify drift using the Bayes Factor of the polynomial relation likelihood change.

cs.LG↗

arXiv2021

Using sequential drift detection to test the API economy

The API economy refers to the widespread integration of API (advanced programming interface) microservices, where software applications can communicate with each other, as a crucial element in business models and functions. The number of possible ways in which such a system could be used is huge. It is thus desirable to monitor the usage patterns and identify when the system is used in a way that was never used before. This provides a warning to the system analysts and they can ensure uninterrupted operation of the system. In this work we analyze both histograms and call graph of API usage to determine if the usage patterns of the system has shifted. We compare the application of nonparametric statistical and Bayesian sequential analysis to the problem. This is done in a way that overcomes the issue of repeated statistical tests and insures statistical significance of the alerts. The technique was simulated and tested and proven effective in detecting the drift in various scenarios. We also mention modifications to the technique to decrease its memory so that it can respond more quickly when the distribution drift occurs at a delay from when monitoring begins.

stat.AP↗

arXiv2021

Density-based interpretable hypercube region partitioning for mixed numeric and categorical data

Consider a structured dataset of features, such as $\{\textrm{SEX}, \textrm{INCOME}, \textrm{RACE}, \textrm{EXPERIENCE}\}$. A user may want to know where in the feature space observations are concentrated, and where it is sparse or empty. The existence of large sparse or empty regions can provide domain knowledge of soft or hard feature constraints (e.g., what is the typical income range, or that it may be unlikely to have a high income with few years of work experience). Also, these can suggest to the user that machine learning (ML) model predictions for data inputs in sparse or empty regions may be unreliable. An interpretable region is a hyper-rectangle, such as $\{\textrm{RACE} \in\{\textrm{Black}, \textrm{White}\}\}\:\&$ $\{10 \leq \:\textrm{EXPERIENCE} \:\leq 13\}$, containing all observations satisfying the constraints; typically, such regions are defined by a small number of features. Our method constructs an observation density-based partition of the observed feature space in the dataset into such regions. It has a number of advantages over others in that it works on features of mixed type (numeric or categorical) in the original domain, and can separate out empty regions as well. As can be seen from visualizations, the resulting partitions accord with spatial groupings that a human eye might identify; the results should thus extend to higher dimensions. We also show some applications of the partition to other data analysis tasks, such as inferring about ML model error, measuring high-dimensional density variability, and causal inference for treatment effect. Many of these applications are made possible by the hyper-rectangular form of the partition regions.

cs.LG↗

arXiv2021

Towards API Testing Across Cloud and Edge

API economy is driving the digital transformation of business applications across the hybrid Cloud and edge environments. For such transformations to succeed, end-to-end testing of the application API composition is required. Testing of API compositions, even in centralized Cloud environments, is challenging as it requires coverage of functional as well as reliability requirements. The combinatorial space of scenarios is huge, e.g., API input parameters, order of API execution, and network faults. Hybrid Cloud and edge environments exacerbate the challenge of API testing due to the need to coordinate test execution across dynamic wide-area networks, possibly across network boundaries. To handle this challenge, we envision a test framework named Distributed Software Test Kit (DSTK). The DSTK leverages Combinatorial Test Design (CTD) to cover the functional requirements and then automatically covers the reliability requirements via under-the-hood closed loop between test execution feedback and AI based search algorithms. In each iteration of the closed loop, the search algorithms generate more reliability test scenarios to be executed next. Specifically, five kinds of reliability tests are envisioned: out-of-order execution of APIs, network delays and faults, API performance and throughput, changes in API call graph patterns, and changes in application topology.

cs.AI↗

arXiv2021

Machine Learning Model Drift Detection Via Weak Data Slices

Detecting drift in performance of Machine Learning (ML) models is an acknowledged challenge. For ML models to become an integral part of business applications it is essential to detect when an ML model drifts away from acceptable operation. However, it is often the case that actual labels are difficult and expensive to get, for example, because they require expert judgment. Therefore, there is a need for methods that detect likely degradation in ML operation without labels. We propose a method that utilizes feature space rules, called data slices, for drift detection. We provide experimental indications that our method is likely to identify that the ML model will likely change in performance, based on changes in the underlying data.

cs.LG↗

arXiv2021

Broadly Applicable Targeted Data Sample Omission Attacks

We introduce a novel clean-label targeted poisoning attack on learning mechanisms. While classical poisoning attacks typically corrupt data via addition, modification and omission, our attack focuses on data omission only. Our attack misclassifies a single, targeted test sample of choice, without manipulating that sample. We demonstrate the effectiveness of omission attacks against a large variety of learners including deep neural networks, SVM and decision trees, using several datasets including MNIST, IMDB and CIFAR. The focus of our attack on data omission only is beneficial as well, as it is simpler to implement and analyze. We show that, with a low attack budget, our attack's success rate is above 80%, and in some cases 100%, for white-box learning. It is systematically above the reference benchmark for black-box learning. For both white-box and black-box cases, changes in model accuracy are negligible, regardless of the specific learner and dataset. We also prove theoretically in a simplified agnostic PAC learning framework that, subject to dataset size and distribution, our omission attack succeeds with high probability against any successful simplified agnostic PAC learner.

cs.LG↗

arXiv2020

A Game Theoretic Model for Strategic Coopetition in Business Networks

Private blockchain is driving the creation of business networks, resulting in the creation of new value or new business models to the enterprises participating in the network. Such business networks form when enterprises come together to derive value through a network which is greater than the value that can be derived solely by any single company. This results in a setting that combines both competitive and cooperative behavior, and which we call strategic coopetition. Traditionally, cooperative and competitive behavior have been analyzed separately in game theory. In this article, we provide a formal model enabling to jointly analyze these different types of behaviors and the interdependencies between them. Using this model, we formally demonstrate and analyze the incentives for both cooperative and competitive behavior.

cs.GT↗

arXiv2020

Sequential Drift Detection in Deep Learning Classifiers

We utilize neural network embeddings to detect data drift by formulating the drift detection within an appropriate sequential decision framework. This enables control of the false alarm rate although the statistical tests are repeatedly applied. Since change detection algorithms naturally face a tradeoff between avoiding false alarms and quick correct detection, we introduce a loss function which evaluates an algorithm's ability to balance these two concerns, and we use it in a series of experiments.

stat.AP↗

arXiv2019

Engineering Reliable Deep Learning Systems

Recent progress in artificial intelligence (AI) using deep learning techniques has triggered its wide-scale use across a broad range of applications. These systems can already perform tasks such as natural language processing of voice and text, visual recognition, question-answering, recommendations and decision support. However, at the current level of maturity, the use of an AI component in mission-critical or safety-critical applications can have unexpected consequences. Consequently, serious concerns about reliability, repeatability, trust, and maintainability of AI applications remain. As AI becomes pervasive despite its shortcomings, more systematic ways of approaching AI software development and certification are needed. These fundamental aspects establish the need for a discipline on "AI Engineering". This paper presents the current perspective of relevant AI engineering concepts and some key challenges that need to be overcome to make significant progress in this important area.

cs.CY↗

arXiv2019

Defending via strategic ML selection

The results of a learning process depend on the input data. There are cases in which an adversary can strategically tamper with the input data to affect the outcome of the learning process. While some datasets are difficult to attack, many others are susceptible to manipulation. A resourceful attacker can tamper with large portions of the dataset and affect them. An attacker can additionally strategically focus on a preferred subset of the attributes in the dataset to maximize the effectiveness of the attack and minimize the resources allocated to data manipulation. In light of this vulnerability, we introduce a solution according to which the defender implements an array of learners, and their activation is performed strategically. The defender computes the (game theoretic) strategy space and accordingly applies a dominant strategy where possible, and a Nash-stable strategy otherwise. In this paper we provide the details of this approach. We analyze Nash equilibrium in such a strategic learning environment, and demonstrate our solution by specific examples.

cs.CR↗

arXiv2016

Towards a Human-Centred Approach in Modelling and Testing of Cyber-Physical Systems

The ability to capture different levels of abstraction in a system model is especially important for remote integration, testing/verification, and manufacturing of cyber-physical systems (CPSs). However, the complexity of modelling and testing of CPSs makes these processes extremely prone to human error. In this paper we present our ongoing work on introducing human-centred considerations into modelling and testing of CPSs, which allow for agile iterative refinement processes of different levels of abstraction when errors are discovered or missing information is completed.

cs.SE↗

arXiv2015

Teaching Logic to Information Systems Students: Challenges and Opportunities

In contrast to Computer Science, where the fundamental role of Logic is widely recognized, it plays a practically non-existent role in Information Systems curricula. In this paper we argue that instead of Logic's exclusion from the IS curriculum, a significant adaptation of the contents, as well as teaching methodologies, is required for an alignment with the needs of IS practitioners. We present our vision for such adaptation and report on concrete steps towards its implementation in the design and teaching of a course for graduate IS students at the University of Haifa. We discuss the course plan and present some data on the students' feedback on the course.

cs.CY↗