Thoughts about talking and seminar
Posted on 2015-09-12 01:31:30 Reading Sharing Views: 695

1. Blueprint for a Science of Security

  • Speaker: Dr. Fred Schneider
  • Time: 2:00PM Sep.11st, 2015
  • Location: IS Building, 3rd Floor collaboration space

To best of our knowledge, in the field of cybersecurity, the defenders have limited resources to defend the system. Because we could develop defense only for attacks that we have already know nowadays. Obviously, the alternative way, or the best way, is to build systems whose security follows from some specific principles. However, unfortunately, we do not know the principles.

That’s the key thoughts of the talking, developing a science of cybersecurity, proposed by the Dr. Fred Schneider. Dr. Fred stand in the high level to think the future of the security in computer science. That is to build the principle of science of cybersecurity.

If that’s the destination of us, the researcher in information security, why not to start earlier? Inspired by the talking, some tips, or maybe method to do research, are given by Dr. Fred. Sometime, it is better to go out of the circle of your facing problem. Stand out of the circle, you may see the essence of the problem and find the way to solve it.

2. The Land Sharks are on the Squawk Box (MIT and Turing Award Winner)

  • Speaker: MICHAEL STONEBRAKER
  • Time: 3:00pm to 4:30pm, Sep 15th, 2015
  • Location: Rashid Auditorium 4401 Gates&Hillman Centers

Michael Stronebraker is the expert in the field of database and also the Turing winner of Turing Award, 2014. The talking was given in CMU. The another title of this report is “How Riding a Bicycle across America and Building Postgres Have a Lot in Common”, which is much more intuitive to understand the report.

The presentation is not much more technical in his specific field of databases. Instead of technical information, Michael shares his experience and attitude on what he was doing in the last years. I think it much helpful to learn something useful than just a technical report.

I still remember the key words, “Make it happen”, even though the talking is 4 days before. Sometime, we all just need a action to start the work, no matter what kind of work. When we start, issues will be found and solved. Michael tell us the story of his bicycle across the U.S. and the history of PostgreSQL. Here is most interesting part I want to share.

Until (ocean) {

get up in the morning;

ride east;

persevere and overcome any obstacles

that arise;

}

make it more universal:

Until(GOAL) {

get up in the morning;

ACTION;

persevere and overcome any obstacles

that arise;

}

So, let’s make it happen.

3. Company Tech Talk: Big Data

  • Speaker: Matt Hunt
  • Time: 5:00pm to 6:30pm, Sep 22nd, 2015
  • Location: Rashid Auditorium 4401 Gates&Hillman Centers

As the big data is a very popular topic in nowadays, I went to the talking when I find the information form CoMeT website. Obviously, it’s just a company tech talk. In general, Matt is a senior developer and the presentation, based on hadoop and big data, discussed the evolving data needs of the finance industry and the scalable solutions they are building to solve.

To be honest, I have few background about finance industry and a few background on hadoop when I works at Baidu Inc. So it’s a little hard for me to understand the talking. However, it is a good opportunity to get some knowledge on finance industry, although I have no much words on finance. It give a glance to know the combination between the finance industry and computer science industry. Beside the technical details, it is a good choice talking for the Tuesday.

4. Thesis Defense: Supporting Learner-Controlled Problem Selection in Intelligent Tutoring Systems

  • Speaker: Yanjin Long
  • Time: 9:00pm to 10:30pm, Sep 25th, 2015
  • Location: GHC 6115

This is a thesis defense of CMU, also the first thesis defense that I attended in the foreign country, the US. To have a listen on the thesis defense is just a coincidence. I went to CMU to meet a friend and find the thesis when I wait for him.

Besides the detail information, which I am not familiar with the concepts(such as self-regulated learning, SRL skills), the content of the defense is in the field of Human Computer Interaction, a new field for me, but it seems interesting. For instance, you need to have design background psychology knowledge and computer science background. I think it’s interdisciplinary.

Another thought I want to show is that the defense procession teaches me a lot. How to defense your point and how to show your work with great presentation skill, both on slides, words and maybe confidence, is my current problems when I have a meeting with advisor and group members. That defense is a intuitive class for me, to some extent.

5. Tachyon: A Memory-centric Distributed Storage System

  • Speaker: Haoyuan LI
  • Time: Oct 01, 2015 6:30 PM - 7:30 PM
  • Location: Panther Hollow Conference Room, 4th Floor Gates&Hillman Centers

As attending many talks before (usually once a week), this is first talking that the speaker is from China. Li has a attracted CV, from Peking University (famous Chinese University) to Cornell University, and then UC Berkeley, which inspires me a lot. Mr. Li is also a founder and CEO of Tachyon Nexus. Thus this is a talking with industrial background.

As data sets continue to grow, storage is increasingly becoming a critical bottleneck in many workloads. To address this problem, Mr. Li’s group developed Tachyon, a memory-centric fault-tolerant distributed storage system, which enables reliable data sharing at memory-speed across cluster frameworks such as Apache Spark, MapReduce, and Apache Flink. The project is also the storage layer of the Berkeley Data Analytics Stack (BDAS) and also part of the Fedora distribution. Thus, in this talk, Mr. Li gives an overview of Tachyon and some instances in the real world.

6. The Lightning Memory-Mapped Databases

  • Speaker: Howard Chu
  • Time: 12:00pm to 1:00pm, Oct. 8th, 2015
  • Location: ISTC Panther Hollow Conference Room, 4th Floor Robert Mehrabian Collaborative Innovation Center

The talking attracted me a lot because of the previous introduction key word, OpenLDAP. As OpenLDAP is an open source project about Lightweight Directory Access Protocol, which lies in my research domains currently, I decide to attend the talking.

The beginning of the talking gives me a surprise for seeing a man with long hair, Asia face, playing violin when I come into the room. It’s a pity that I can not play instruments though I also love music. Back to topic, the main idea is about the LMDB, Lightning Memory Mapped DB which is designed for the OpenLDAP project. Mr. Chu introduce motivation (The caches were not always beneficial and were sometimes detrimental), comparison about the DB they are working on. He gives some basic idea about the LMDB, such that Cache management is a hassle, so do not do any caching; lock management is a hassle, so do not do any locking.

Although the talking is not about the LDAP, it also gives me some insight on the DB field, which lies in my interesting topics. Sometime, we should focus on specific problem and gives daring solution to that. Maybe the result will beyond your expectation.

7. The Open Movement in Higher Education

  • Speaker: Sheila Corrall
  • Time: Oct 15th, 2015 10:30 PM - 12:00 PM
  • Location: University of Pittsburgh Hillman Library, Amy Knapp Room

As so many technical talks in these days, this Thursday, I attended a talk in the Library, just because I were here. The topic is about the Open Movement in Higher Education, which focus on the range and state of open activities in the higher education area, some common factors for the different open domains and potential benefits for individuals. Also, the talk gives idea about opportunities and challenges, but I do not fully understand.

Bringing the topic to a specific field, such as computer science, that I familiar with, I can think of the open source project. There are many open source project online and open source organizations. I think that is also open movement in computer science. For me, I prefer the open source project, which gives us a platform to do implementation in one field and to get a better tool with cooperation all over the world. Thus open is a kind of attitude. To some extend, the publishing in the research field is also a open movement. You can get more comments, more challenges on your work, which is better for you to modify.

8. Research at Microsoft: Looking Beyond the Horizon

  • Speaker: JEANNETTE M. WING
  • Time: Thursday, October 22, 2015, 4:00pm to 5:00pm
  • Location: Rashid Auditorium 4401 Gates&Hillman Centers

As we know, the Microsoft Research is an good research institute over the word. For me, it is an idea research place in the industrial fields. As an Chinese students, Microsoft Research Asia (MSRA) is a wonderful place to do research because of its ideal working environment and better welfare. Also, it is a good choice for my career in future if I go back to China when graduated. Thus it is an opportunity to have a glance to research at Microsoft.

Mrs Wing is Corporate Vice President, Microsoft Research. She was Professor of Computer Science at Carnegie Mellon (and is currently adjunct faculty) and twice served as the CMU Head of the Computer Science Department.

This talk gives me new discovery for Microsoft Research. Microsoft Research is a unique industrial research lab supporting open, basic research with a mission to advance state of the art while doing rapid technology transfer to industry. Besides, Mrs Wing gives me lots of information about the culture and values in the MSR, which seems to be an important part to them leading to high-caliber research.

9. Achieving Cybersecurity Using Semantic Diversity

  • Speaker: Dr. Stuart Rubin
  • Time: October 30, 2015, 2:00pm to 3:30pm
  • Location: IS Building, 3rd Floor collaboration space

This is a seminar related to our lab this week. Dr. Stuart Rubin is a senior scientist of SPAWAR. As we know, information reuse and integration is needed to provide military forces with information dominance. This implies protecting our ever-more complex software systems from infiltration. This, in turn, requires higher-level compilers to make semantic diversity cost effective.

So, in this talk, Dr. Rubi encompasses some topics related, such that the major present approaches to cybersecurity based on diversity; applying randomization to cybersecurity. Also, there are some topic about semantic randomization theorem. For me, some of them is new and they gives me a glances on the field between security and data analysis. I think it is a good direction to my further work.

10. Building a Network Security Analysis Toolbox: Problems of Vantage, Narrative and Action

  • Speaker: MICHAEL COLLINS
  • Time: Nov 2ed, 2015, 12:00pm to 1:00pm
  • Location: Distributed Education Classroom

It’s a talk about the information security area, which is similar to my current research field. Thus, the seminar of attracts me a lot. Michael Collins is the chief scientist of RedJack, LLC that provides government agencies and corporations with targeted solutions for complex enterprise network security problems. In the company, Michael has led research on insider threat, game-theoretic security, moving target defense and software defined networking.

In the field of information security, the researcher will face the security problem found by the attackers. It’s necessary to transfer research into operations. The talking is about how to build a network security analysis toolbox which takes security research and turns it into actionable analysis on networks that comprise millions of IP address.

For me, the technical details is not key part of the talk. I appreciate that the methodology. As a researcher, when doing some research work, we also need to make our own toolbox, which could be a simple python code to collect data, or check the confidentiality or availability of the data. Although we need much more time to make the tool to be more universal for future using, the future research work will benefit a lot from the toolbox. Maybe save a lot of time.

11.Securing the Future

  • Speaker: Darren Shou
  • Time: Nov 9th, 2015, 12:00pm to 1:00pm
  • Location: Distributed Education Classroom

As Mr. Shou is the director of Symantec Research Labs Core Research Team, which is a famous security company in the industry, he has much experience on the engineering of the field. Also, he serves on the industry advisory boards of the National Science Foundation TRUST program and the Georgia Tech Information Security Center.

For his bi-background in industry and academia, the talk is in a high level about the security of computer science which includes several of shifts about potential disruptions on the horizon that impacts the nature of business and the structure of industries. Also, the talk includes unique challenges and opportunities for the security domain.

Form the several disruptive innovations, such as 3D printing, Mr. Shou takes it as an example, I thinks the security will be a standalone feature of any new technologies. No matter what the technology is, there are still attackers or adversaries want to use them for personal purposes. Thus when we developed a new technology, we should consider the security side or property of the technology first. Maybe we need a universal security framework to secure the future.

12.PhD Dissertation Defense: Techniques for Application-Aware Suitability Analysis of Access Control Systems

  • Speaker: William C. Garrison III
  • Time: Nov 13, 2015, 1:30pm to 3:00pm
  • Location: 6329 SENSQ

This is the dissertation defense in department of computer science and the topic is all about my current research topic, access control. This work was not how to do the access control framework, but to analysis of a access control system.

Prior work on evaluating access control systems is best has primarily relied on relative expressiveness analysis, then proves that one system has greater capabilities than another. But there still some issues, such that the applications in the system, ranking way, and so on.

The thesis of William was to address such kind of shortcomings described above. He propose the access control suitability analysis problem, and present a series of techniques for solving it. Moreover, his framework was evaluated formally by proving that is satisfies a series of technical requirements and practically by presenting several case studies in realistic scenarios.

13.What You Need to Know about Chinese for Chinese Language Processing

  • Speaker: Chu-ren Huang
  • Time: Nov 23, 2015, 12:00pm to 1:00pm
  • Location: 8102 Gates&Hillman Centers

Obviously, its topic is all about the field of natural language processing (NLP). Dr. Huang is from the Hong Kong Polytechnic University, who is the chair professor of applied Chinese Language Studies.

In this talk, essential knowledge of Chinese linguistics which includes both the fundamental knowledge of the linguistic structure and explanations was introduced. Aiming to provide comprehensive knowledge of the linguistic characteristics of the Chinese language, the perspective is synergetic.

As Chinese is my native language, it’s fresh that the talk gives some new thoughts about technical ways to re-recognize my native language. For example, the characters, words, speech with sentence and structure. How to identify grammatical relations in Chinese as well as a few Chinese-specific constructions.

14.Online Privacy Expectations: Types, Mismatches and Implications for Privacy Notices

  • Speaker: Ashwini Rao
  • Time: Nov 30, 2015, 2:30pm to 3:30pm
  • Location: ASA Conference Room 6115 Gates&Hillman Centers

This is a thesis proposal of Ph.D. student, Ashwini Rao, from software engineering, institute for software research. As the online privacy notices is a quite simple and primary mechanism to inform users about data issues. However, usually, users always ignore such kind of notices for some reasons, maybe too long and complex to read.

The proposal was to simplify privacy notices by understanding mismatches between users’ privacy expectations regarding data practices of online services and actual data practices of online services. Thus, he try to present an approach for identifying such mismatches.

For me, I think it’s novel, but my concerns is that how to evaluate the result. Obviously, the privacy notice is shown everyday, different users will have different view to them. When you mentioned your approach is quite simple, it depend on users’ subjective view. The evaluation part of the proposal gives me some new view on such kind of issues that I had before. Maybe that is the biggest apple I got from the proposal.

15.Helping Users Manager their Privacy through Nudges

  • Speaker: Hazim Almuhimedi
  • Time: Dec 2, 2015, 3:30pm to 4:30pm
  • Location: Reddy Conference Room 4405 Gates&Hillman Centers

As the talk is a related topic in security field and hold on Wednesday, I just get last half part of the talk because of the time conflict to my class. It seems that its applying domain is on apps in the smartphones. The speaker want to help users to manage their privacy on mobile devices because, obviously, privacy issues are important in our life by using smart phones.

Thus the speaker propose to adopt the concept of nudges, a kind of soft paternalistic behavioral interventions. It helps users make better privacy decisions by

(1) increasing users’ awareness about privacy risks associated with apps that users have previously installed,
(2) motivating users to review their app privacy settings and adjust them as needed.

For me, the work may be not quite academic but more about engineering. And then I found that it is also a proposal from Institute for Software Research, just like the previous I attends several days ago. Anyway, it’s also worthy to attend. Life is a way to find the problems, no matter how big it is, and then solve it by our approaches.