David Wagner

University of New Brunswick

Author Statistics

Papers

Citation

H-Index

i-10 index

Research Trends

Author Order

Document Type

Co-Authors

Serge Egelman

International Computer Science Institute

Chawin Sitawarin

University of California, Berkeley

Beth Herbel‐Eisenmann

Michigan State University

Nicholas Carlini

Google (United States)

Jennifer Barton Gilman

Annica Andersson

University of South-Eastern Norway

Adrienne Porter Felt

Google (United States)

Primal Wijesekera

International Computer Science Institute

Konstantin Beznosov

University of British Columbia

Erika Chin

Pushkin Museum

Cooperative Institutions

University of California, Berkeley

124

Berkeley College

Google (United States)

Stanford University

Princeton University

University of California System

University of California, San Diego

Carnegie Mellon University

University of Illinois Urbana-Champaign

Massachusetts Institute of Technology

Author Statistics

Papers

Citation

H-Index

i-10 index

Research Field

Respond or Dismiss: Interactions That May Support Loving, Bullying and Solitude in Mathematics

Journal of the Canadian Association for Curriculum Studies (2019)

Annica Andersson David Wagner

We develop an analytical frame to support reflection on how love, bullying and solitude appear in communication in mathematics classrooms. The frame distinguishes among communication acts that are responsive or dismissive to others, and identifies how the acts draw on authority to open or close dialogue. Our examination of high school students’ language repertoires revealed that their communication acts influence both their development of mathematics and their interpersonal relationships. To illustrate the framework in use, we draw on transcripts from group interactions in one high school mathematics class. We consider how the particular kinds of communication acts may support and develop caring, antagonism or reclusion. The illustrative analysis illuminates the complexity of human relationships—how communication acts in mathematics classrooms intertwine personal autonomy, interpersonal positioning, and how these communication acts are intertwined with interpersonal positioning.

Solitude

10.25071/1916-4467.40399

Cite

Citations (3)

Aiming for Equity in Ethnomathematics Research

Springer eBooks (2011)

David Wagner Lisa Borden

Ethnomathematics

CONTEST

Equity

10.1007/978-94-007-2813-4_5

Cite

Citations (6)

Jatmo: Prompt Injection Defense by Task-Specific Finetuning

arXiv (Cornell University) (2023)

Julien Piet Maha Al-Rashed Chawin Sitawarin Sizhe Chen Zeming Wei

Large Language Models (LLMs) are attracting significant research attention due to their instruction-following abilities, allowing users and developers to leverage LLMs for a variety of tasks. However, LLMs are vulnerable to prompt-injection attacks: a class of attacks that hijack the model's instruction-following abilities, changing responses to prompts to undesired, possibly malicious ones. In this work, we introduce Jatmo, a method for generating task-specific models resilient to prompt-injection attacks. Jatmo leverages the fact that LLMs can only follow instructions once they have undergone instruction tuning. It harnesses a teacher instruction-tuned model to generate a task-specific dataset, which is then used to fine-tune a base model (i.e., a non-instruction-tuned model). Jatmo only needs a task prompt and a dataset of inputs for the task: it uses the teacher model to generate outputs. For situations with no pre-existing datasets, Jatmo can use a single example, or in some cases none at all, to produce a fully synthetic dataset. Our experiments on seven tasks show that Jatmo models provide similar quality of outputs on their specific task as standard LLMs, while being resilient to prompt injections. The best attacks succeeded in less than 0.5% of cases against our models, versus 87% success rate against GPT-3.5-Turbo. We release Jatmo at https://github.com/wagner-group/prompt-injection-defense.

Leverage (statistics)

10.48550/arxiv.2312.17673

Cite

Citations (0)

Situated Mathematics: Positioning Mathematics Ideas as Human Ideas

WORLD SCIENTIFIC eBooks (2019)

David Wagner

10.1142/9789811205385_0004

Cite

Citations (2)

Part-Based Models Improve Adversarial Robustness

arXiv (Cornell University) (2022)

Chawin Sitawarin Kornrapat Pongmala Yizheng Chen Nicholas Carlini David Wagner

We show that combining human prior knowledge with end-to-end learning can improve the robustness of deep neural networks by introducing a part-based model for object classification. We believe that the richer form of annotation helps guide neural networks to learn more robust features without requiring more samples or larger models. Our model combines a part segmentation model with a tiny classifier and is trained end-to-end to simultaneously segment objects into parts and then classify the segmented object. Empirically, our part-based models achieve both higher accuracy and higher adversarial robustness than a ResNet-50 baseline on all three datasets. For instance, the clean accuracy of our part models is up to 15 percentage points higher than the baseline's, given the same level of robustness. Our experiments indicate that these models also reduce texture bias and yield better robustness against common corruptions and spurious correlations. The code is publicly available at https://github.com/chawins/adv-part-model.

Robustness

Spurious relationship

End-to-end principle

10.48550/arxiv.2209.09117

Cite

Citations (1)

Adversarial Examples for k-Nearest Neighbor Classifiers Based on Higher-Order Voronoi Diagrams

Neural Information Processing Systems (2021)

Chawin Sitawarin Evgenios M. Kornaropoulos Dawn Song David Wagner

Adversarial examples are a widely studied phenomenon in machine learning models. While most of the attention has been focused on neural networks, other practical models also suffer from this issue. In this work, we propose an algorithm for evaluating the adversarial robustness of $k$-nearest neighbor classification, i.e., finding a minimum-norm adversarial example. Diverging from previous proposals, we take a geometric approach by performing a search that expands outwards from a given input point. On a high level, the search radius expands to the nearby Voronoi cells until we find a cell that classifies differently from the input point. To scale the algorithm to a large $k$, we introduce approximation steps that find perturbations with smaller norm, compared to the baselines, in a variety of datasets. Furthermore, we analyze the structural properties of a dataset where our approach outperforms the competition.

Robustness

Source

Cite

Citations (0)

Detecting and Characterizing Lateral Phishing at Scale

arXiv (Cornell University) (2019)

Grant Ho Asaf Cidon Lior Gavish Marco Schweighauser Vern Paxson

We present the first large-scale characterization of lateral phishing attacks, based on a dataset of 113 million employee-sent emails from 92 enterprise organizations. In a lateral phishing attack, adversaries leverage a compromised enterprise account to send phishing emails to other users, benefitting from both the implicit trust and the information in the hijacked user's account. We develop a classifier that finds hundreds of real-world lateral phishing emails, while generating under four false positives per every one-million employee-sent emails. Drawing on the attacks we detect, as well as a corpus of user-reported incidents, we quantify the scale of lateral phishing, identify several thematic content and recipient targeting strategies that attackers follow, illuminate two types of sophisticated behaviors that attackers exhibit, and estimate the success rate of these attacks. Collectively, these results expand our mental models of the 'enterprise attacker' and shed light on the current state of enterprise phishing attacks.

Phishing

Leverage (statistics)

True positive rate

10.48550/arxiv.1910.00790

Cite

Citations (20)

The effect of developer-specified explanations for permission requests on smartphone user behavior

Joshua Tan Khanh Nguyen Michael Theodorides Heidi Negrón-Arroyo Christopher Thompson

In Apple's iOS 6, when an app requires access to a protected resource (e.g., location or photos), the user is prompted with a permission request that she can allow or deny. These permission request dialogs include space for developers to optionally include strings of text to explain to the user why access to the resource is needed. We examine how app developers are using this mechanism and the effect that it has on user behavior. Through an online survey of 772 smartphone users, we show that permission requests that include explanations are significantly more likely to be approved. At the same time, our analysis of 4,400 iOS apps shows that the adoption rate of this feature by developers is relatively small: around 19% of permission requests include developer-specified explanations. Finally, we surveyed 30 iOS developers to better understand why they do or do not use this feature.

Permission

Feature (linguistics)

10.1145/2556288.2557400

Cite

Citations (58)

Students' Critical Awareness of Voice and Agency in Mathematics Classroom Discourse

Mathematical Thinking and Learning (2007)

David Wagner

10.1207/s15327833mtl0901_3

Cite

Citations (20)

A Large-Scale Analysis of Attacker Activity in Compromised Enterprise Accounts

Communications in computer and information science (2020)

Neil Shah Grant Ho Marco Schweighauser Mohamed Ibrahim Asaf Cidon

10.1007/978-3-030-59621-7_1

Cite

Citations (1)