Events

Speaker:  Dr. Edward A. Lee, Professor, University of California at Berkeley

Title:  Deterministic Concurrency for Cyber-Physical Systems

Abstract/Bio:  TBA

Host:  Arpan Gujarati, UBC Computer Science

As a way to reconnect and build community among CS alumni, the Computer Science Department is offering a Machine Learning workshop on Saturday, Dec 7, from 9:30 am to 4 pm in Rm 4074 of the UBC Orchard Commons.

In this workshop, we will demystify the fundamentals of machine learning and clarify what it can and cannot do for your projects. By the end of the workshop, you’ll have a solid understanding of machine learning fundamentals and different types of machine learning. You’ll also get hands-on practice developing simple machine learning pipelines.

Please sign up below if you are interested in coming.  A reminder will be send out a week before the event.  Any cancellation of this initiative will be shared via email 2 working days prior.

*Lunch is not included, there are multiple delicious food options available across campus.

Agenda:

9:30 to 10:00 | Introduction & objectives 

10:00 to 10:30 | Machine learning fundamentals 

10 to 11:00 | Hands-on activity 1: Data exploration and preparation 

11:00 to 11:15 | Break 

11:15 to 12:00 | Supervised learning

12:00 to 12:30 | Hands-on activity 2: Building a simple classification model

12:30 to 1:30 | Lunch Break

1:30 to 2:30 | Unsupervised Learning: Clustering and dimensionality reduction 

2:30 to 3:30 | Hands-on activity 2: Building a simple classification model

3:30 to 4:00 | Closing remarks and Q&A

Questions?  Please email Michele Ng at mng@cs.ubc.ca

Name: Ishita Haque 

Date: Nov 27, 2024

Time: 2-3 PM

Location: x836

Supervisor: Dr. Joanna McGrenere

Title: Envisioning Interventions to Combat Misinformation Propagation on Social Media: Insights from Older Adults’ Approaches to Credibility Assessment and Sharing Decisions

Abstract:
Users of all ages contribute to misinformation propagation on social media, including older adults. With the growing adoption of social media among older adults, it is essential to understand their perspectives on credibility assessment and its perceived impact on sharing decisions.

Leveraging friends and family (FnFs) to support credibility assessment is a promising approach; however, little is known about how older adults value it relative to assessing credibility individually.

To probe this question, we created a prototype that nudges before sharing potential misinformation and engaged 12 older adults in think-aloud sessions and semi-structured interviews over selected social media posts. Our thematic analysis reveals that older adults prefer an independent approach while assessing credibility, but that involving FnFs is often desired for social opportunities for shared decision-making and enhancing interpersonal relationships. Older adults perceive multifaceted risks when sharing across inner (FnFs) and outer circles, and they exhibit varying trust in FnFs’ expertise and individually sourced information. They also attempt to mitigate their prior beliefs through cross-checking. We offer a nuanced understanding of older adults’ perceptions towards credibility and sharing and highlight the design challenges and opportunities for empowering older adults and mitigating social risks in seeking assessment support.

Name: Wanxin Li

Date: Nov 26, 2024,

Time: 2 pm - 4 pm

Location: ICICS 146

Supervisors: Anne Condon,  Khanh Dao Duc

Title: Computational Methods for Addressing Bias and Fairness and Analyzing Cell Shape Heterogeneity

Abstract:

This thesis proposal focuses on two problems arising from biomedical data: (i) Addressing bias and fairness in healthcare systems and (ii) Improving metric and dimensionality reduction in cell shape heterogeneity analysis. 

Part (i): As healthcare systems adopt Electronic Health Records (EHRs), more opportunities arise for algorithmic applications. However, two challenges arise: First, EHRs from different populations can introduce biases, making data and models less transferable. To address this, we developed a method leveraging optimal transport (OT) to transfer knowledge between populations with guarantees for suitability and enabling the quantification of treatment disparities. Second, algorithms using EHRs may propagate unfairness. While fairness testing methods have been proposed for binary classification, they often lack computational tractability. We aim to propose a framework based on OT projections for testing the fairness of regression problems under a wide range of fairness criteria while maintaining computational tractability. We will apply the framework to test the fairness of critical algorithms in healthcare, such as emergency room wait time prediction. 

Part (ii): Cells grown on planar surfaces show diverse morphological shapes due to genetic or environmental factors. We explored how alternative metrics, beyond the Euclidean metric, can offer insights into cell shape heterogeneity. The Square Root Velocity (SRV) metric, a specific instance of the elastic metric, is known for its computational efficiency in practice. While current studies restricted the use of the SRV metric to either simple shapes or basic tasks, we extensively explored SRV's power in comparing its distances to the mean shape and using it as the metric for multi-dimensional scaling (MDS). Our study showed the superior performance of SRV against the linear metric on datasets of human cancer cells. Additionally, the presence of orthogonal outliers can significantly distort results from both Euclidean and non-Euclidean metrics. To address this, we developed a method that detects and corrects orthogonal outliers in MDS while estimating the dataset's dimensionality. We validated the effectiveness of this method on single-cell images, microbiome sequencing data, and scRNA-seq data.

Speaker:  Stefan Saroiu, Senior Principal Researcher, Microsoft Research

Title:  Six Years of Rowhammer: Breakthroughs and Future Directions

Abstract:

This talk will present the work done over the past six years as part of Project STEMA at Microsoft. STEMA stands for Secure, Trusted, and Enhanced Memory for Azure. We will discuss our journey in understanding Rowhammer and our methodology for determining whether cloud servers are vulnerable to these attacks. We will also explain why Rowhammer is a significant concern, particularly in the context of nation-state attacks, and how this led us to develop a pragmatic solution called Panopticon.

We will then introduce Panopticon, an in-DRAM Rowhammer defense that is cost-effective and requires no hardware changes beyond DRAM itself. Unlike previous solutions that monitor Rowhammer in SRAM or CAM memories, Panopticon is the first to implement monitoring within the DRAM fabric. Combined with its alert system, Panopticon has the potential to address Rowhammer once and for all.

Panopticon's approach has caught the attention of industry, leading to the development of Per-Row Activation Counting (PRAC), a groundbreaking Rowhammer defense that will soon be widely deployed in most, if not all, DRAM. In the final part of our talk, we will do a brief technical deep dive into PRAC. While PRAC marks a significant advance in DRAM security, its specification leaves some questions unanswered and exposes potential gaps and challenges. This presents a huge opportunity for the research community to address these issues.

Bio:

Stefan Saroiu is a researcher with Microsoft Research. His research interests cover many aspects of systems and networks, although his recent work has primarily focused on systems security. Stefan's work has been published at top conferences in the fields of security, systems, networking, and mobile computing.

Stefan takes his work beyond publishing results. With his colleagues at Microsoft, he (1) designed Panopticon, a Rowhammer defense adopted by the DRAM industry, (2) designed, deployed, and operated Microsoft Embedded Social, a cloud service aimed at user engagement in mobile apps, which had 20 million users, (3) created the reference implementation of a software-based Trusted Platform Module (TPM) used in hundreds of millions of smartphones and tablets, and (4) designed and operated Zero-Effort Payments (ZEP), one of the first face recognition-based payment systems in the world.

Before joining Microsoft in 2008, Stefan spent three years as an Assistant Professor at the University of Toronto, and four months as a visiting researcher at Amazon.com, where he contributed to the early designs of their new shopping cart system, also known as Dynamo. Stefan has a PhD from the University of Washington and is an ACM Fellow.

Host:  Aastha Mehta, UBC Computer Science