arXiv subjects

Aditya Saligrama

Publications and source records attributed to Aditya Saligrama.

5 recordsLinked to original sources

arXiv2024

Teaching Cloud Infrastructure and Scalable Application Deployment in an Undergraduate Computer Science Program

Making successful use of cloud computing requires nuanced approaches to both system design and deployment methodology, involving reasoning about the elasticity, cost, and security models of cloud services. Building cloud-native applications without a firm understanding of the fundamentals of cloud engineering can leave students susceptible to cost and security pitfalls. Yet, cloud computing is not commonly taught at the undergraduate level. To address this gap, we designed an undergraduate-level course that frames cloud infrastructure deployment as a software engineering practice. Our course featured a number of hands-on assignments that gave students experience with modern, best-practice concepts and tools including infrastructure-as-code (IaC). We describe the design of the course, our experience teaching its initial offering, and provide our reflections on what worked well and potential areas for improvement. Our course material is available at https://infracourse.cloud.

cs.CY↗

arXiv2020

Revisiting Ensembles in an Adversarial Context: Improving Natural Accuracy

A necessary characteristic for the deployment of deep learning models in real world applications is resistance to small adversarial perturbations while maintaining accuracy on non-malicious inputs. While robust training provides models that exhibit better adversarial accuracy than standard models, there is still a significant gap in natural accuracy between robust and non-robust models which we aim to bridge. We consider a number of ensemble methods designed to mitigate this performance difference. Our key insight is that model trained to withstand small attacks, when ensembled, can often withstand significantly larger attacks, and this concept can in turn be leveraged to optimize natural accuracy. We consider two schemes, one that combines predictions from several randomly initialized robust models, and the other that fuses features from robust and standard models.

stat.ML↗

arXiv2019

KnowBias: Detecting Political Polarity in Long Text Content

We introduce a classification scheme for detecting political bias in long text content such as newspaper opinion articles. Obtaining long text data and annotations at sufficient scale for training is difficult, but it is relatively easy to extract political polarity from tweets through their authorship. We train on tweets and perform inference on articles. Universal sentence encoders and other existing methods that aim to address this domain-adaptation scenario deliver inaccurate and inconsistent predictions on articles, which we show is due to a difference in opinion concentration between tweets and articles. We propose a two-step classification scheme that uses a neutral detector trained on tweets to remove neutral sentences from articles in order to align opinion concentration and therefore improve accuracy on that domain. Our implementation is available for public use at https://knowbias.ml.

cs.CL↗

arXiv2019

KnowBias: A Novel AI Method to Detect Polarity in Online Content

We propose a novel training and inference method for detecting political bias in long text content such as newspaper opinion articles. Obtaining long text data and annotations at sufficient scale for training is difficult, but it is relatively easy to extract political polarity from tweets through their authorship; as such, we train on tweets and perform inference on articles. Universal sentence encoders and other existing methods that aim to address this domain-adaptation scenario deliver inaccurate and inconsistent predictions on articles, which we show is due to a difference in opinion concentration between tweets and articles. We propose a two-step classification scheme that utilizes a neutral detector trained on tweets to remove neutral sentences from articles in order to align opinion concentration and therefore improve accuracy on that domain. We evaluate our two-step approach using a variety of test suites, including a set of tweets and long-form articles where annotations were crowd-sourced to decrease label noise, measuring accuracy and Spearman-rho rank correlation. In practice, KnowBias achieves a high accuracy of 86 (rho = 0.65) on these tweets and 75 (rho = 0.69) on long-form articles. While we validate our method on political bias, our scheme is general and can be readily applied to other settings, where there exist such domain mismatches between source and target domains. Our implementation is available for public use at https://knowbias.ml.

cs.CL↗

arXiv2019

A Practical Analysis of Rust's Concurrency Story

Correct concurrent programs are difficult to write; when multiple threads mutate shared data, they may lose writes, corrupt data, or produce erratic program behavior. While many of the data-race issues with concurrency can be avoided by the placing of locks throughout the code, these often serialize program execution, and can significantly slow down performance-critical applications. Programmers also make mistakes, and often forget locks in less-executed code paths, which leads to programs that misbehave only in rare situations. Rust is a recent programming language from Mozilla that attempts to solve these intertwined issues by detecting data-races at compile time. Rust's type system encodes a data-structure's ability to be shared between threads in the type system, which in turn allows the compiler to reject programs where threads directly mutate shared state without locks or other protection mechanisms. In this work, we examine how this aspect of Rust's type system impacts the development and refinement of a concurrent data structure, as well as its ability to adapt to situations where correctness is guaranteed by lower-level invariants (e.g., in lock-free algorithms) that are not directly expressible in the type system itself. We detail the implementation of a concurrent lock-free hashmap in order to describe these traits of the Rust language. Our code is publicly available at https://github.com/saligrama/concache and is one of the fastest concurrent hashmaps for the Rust language, which leads to mitigating bottlenecks in concurrent programs.

cs.DC↗