Mihai Avram
Hello,

I am a first-year Ph.D. student studying Information Science at the University of Illinois at Urbana-Champaign under Dr. Jana Diesner. My current research interests span Data Science for Social Good, Games With a Purpose, and Impact Assessment. More specifically, I am the creator of Fakey (a news literacy game) and am currently focusing on adversarial learning in social networks as well as uncovering social media misinformation patterns. In the past, I have worked on Hoaxy (a misinformation visualization tool), link prediction on knowledge networks, graph counting, and some IT consulting work.

What am I up to nowadays?

1) Developing methodology and applied experimentation with Adversarial Learning on Social Networks
2) Mining for patterns in how people interact with information/misinformation/disinformation on social media feeds using Fakey
3) Taking courses:
   Social Spaces on the Internet (CS 598) by Karrie Karahalios
   Research Design in LIS (LIS 588) by Jana Diesner
   Adv. Topics in Machine Learning & Social Computing (IS 590) by Jana Diesner (audit)
   Seminar Series (discuss seminal papers in the field)
4) Continuing to expand Fakey in different countries as well as adding extra features
5) Brainstorming ways to work on Impact Assessment and Global Priorities Research as I believe this topic to be a neglected area of research for our time

Resume: Download As PDF
Publications:

HarpLDA+: Optimizing Latent Dirichlet Allocation for Parallel Efficiency
Bo Peng, Bingjing Zhang, Langshi Chen, Mihai Avram, Robert Henschel, Craig Stewart, Shaojuan Zhu, Emily Mccallum, Lisa Smith, Tom Zahniser, Jon Omer, Judy Qiu
Conference - IEEE Big Data 2017
(Download as PDF)

Finding and counting tree-like subgraphs using MapReduce
Zhao Zhao, Langshi Chen, Mihai Avram, Meng Li, Guanying Wang, Ali Butt, Maleq Khan, Madhav Marathe, Judy Qiu, Anil Vullikanti
Journal - IEEE Transactions on Multi-Scale Computing Systems 2017
(Download as PDF)

RelSifter: Scoring Triples from Type-like Relations
Prashant Shiralkar, Mihai Avram, Giovanni Luca Ciampaglia, Filippo Menczer, Alessandro Flammini
Conference - WSDM Cup 2017
(Download as PDF)

A Fast Algorithm to the Radiative Transport Equation and Implementation of Theory Into an Applet
Mihai Avram, Manabu Machida
Report - 2013
(Download as PDF)

Civilian Deaths and the Iraq War
Mihai Avram, Sorin Matei
Journal - Journal of Purdue Undergraduate Research 2013
(Download as PDF)
Current Projects
Adversarial Learning on Social Networks
In this project we are aiming to find the optimal adversarial sequence which can yield positive or negative outcomes in a social network. For instance, if a person has very few friends and wants to be very influential in a short amount of time, what are the most optimal moves that the given person should make in order to achieve the maximum amount of influence.
Fakey - The News Feed Literacy Game (Web, Android , iOS)
Fakey is a game with a purpose where players are placed in a simulated social media news feed and are tasked with disambiguating between high-quality news and lower quality news such as fake, bias, conspiracy, junk science, etc. The players are constructively directed towards creating a good social media experience where good information is shared and false information is fact-checked. We created this game for media literacy as well as for research in uncovering patterns in the way people interact with social media news feeds.
Social Impact Assessment
Thinking of novel ways to measure Social Impact particularly using Games With a Purpose or Social Sensing, and/or Information Retrieval Complex Systems. Currently, this project is at the literature review and brainstorming phase.
Past Projects
Hoaxy - The Social Media Information/Misinformation Diffusion Tool (Web)
Hoaxy is a Social Media visualization tool which can capture virtually any story online, track it back to the accounts that are chatting about it, as well as the level of automation (are the accounts more human like or bot like) in the agents that are spreading the stories. The stories can be real news, fake news, and any spectrum in between.
Graph Counting Using High-Performance Computing
Our approach was to take very large graphs which contained 300M+ nodes and 1B+ edges and count the number of embeddings (sequences of DNA for instance, or number of single-child families). In order to perform such a feat, we used an HPC Machine Learning framework called Harp, and implemented various algorithms that could scale to many nodes and edges using a high-performance computing setup with many nodes, cores, threads, and shared memory.
Relational Based Learning on a Knowledge Base
In this project we looked at WikiData and DBpedia, two large knowlede bases, and attempted to answer questions for entities that did not have links. For instance, what nationality was Barack Obama, or in what profession was Elon Musk using Machine Learning and Network Science approaches.
Echo Global Logistics Job Board (Web)
During my time as a consultant working for Avanade, I was tasked with web development component of the job board for Echo Global Logistics.
Federal Home Loan Bank of Chicago (Dev-Ops Role) (Web)
During my time as a consultant working for Avanade, I was part of the Dev-Ops team for nearly two years where I have taken part of the development lifecycle using the .NET stack and have supported many critical failures, live launches, as well as implemented novel features for the bank in order to enable it to function more efficiently.
EZ-RTE - Creating an Applet for Radiative Transport in 3D (Undergraduate REU)
During the summer of 2013, I worked with Physics professor Manabu Machida in creating an applet for the Radiative Transport phenomenon in three dimensions. Dr. Machida, taught me the theory behind the phenomenon, and I was tasked with creating an applet that can visualize and graph the intensity of light based on the constants of absorption and scattering as light protrudes random media. The applet was created in Java and can be found here.
Civilian Deaths and the Iraq War (Undergraduate Project)
In my first foray into research, my adviser and I looked at civilian deaths and the iraq war using the Wiki-Leaks war logs dataset in order to uncover some of the entities responsible for the deaths of civilians.

Summer 2018

Accomplished:
Took the Pro-Truth Pledge. Visited my beautiful homeland Romania. Started rotating with the new lab, I will be part of at UIUC where I will start my Ph.D. in the coming fall. Working on publishing two papers on Adversarial Learning in Social Networks as well as Patterns of Social Interaction in Social Media feeds with respect to Misinformation and Percieved Engagement.
Lessons Learned: The practice of minimizing the nonessential and focusing time on the essential at a macro and micro level (hence I minimized social media use and quit using Snapchat for once!). Social impact is my harmonious passion. I often think about the best ways to achieve the highest social impact in the world as possible. Not because I want money, glory, or fame, but because I think it our responsability to give back and leave the world a little better off when we die. This is related to the concept of Effective Altruism. Hence, I am constantly thinking about the optimal way to do this whether it is through academia, entrepreneurship, both, or some other way and I hope to figure it out soon!

Spring 2018

Accomplished:
Completed my Master's thesis titled "Hoaxy and Fakey: Tools to Analyze and Mitigate the Spread of Misinformation in Social Media" and launched Fakey for Android and iOS. Also assisted in the launch of a new version of Hoaxy.
Lessons Learned:
Learned to use Apache Cordova in being able to port website code into iOS/Android applications seemlessly and having the end result feel very native to the platform environment. Creating tools for the social good takes a lot of hard work, dedication, and sacrifice - especially without a large support group in assistance. As demanding as research/academia/entrepreneurship may be, social connections and friendships should never be neglected as that can impact one's mental state in the present and future in a negative way.

Fall 2017

Accomplished:
Learned about Natural Language Processing and lauched an alpha version of Fakey , a media literacy game.
Lessons Learned:
Learned the Vue front-end framework, and read up on many Social Computing papers to understand the diverse research done in this field. Hopefully this literature review will better allow me to pick certain sub-areas of research in order to have the most impact and fun in the future. I also learned to find a good work/life balance for myself in the academic environment.

Summer 2017

Accomplished:
Worked for professors Judy Qiu and Ying Ding and published two papers as well as one Machine Learning MOOC.
Lessons Learned:
Focus on the essential and do your best, as trying to do e verything is the recipe for failure. Honed Machine Learning, Graph Algorithms, Linux, and High-Performance Computing skills.

Spring 2017

Accomplished:
Completed a Cloud Computing course and Data Mining Course as well as started formally working with professor Filippo Menczer who turned out to be my Master's research advisor.
Lessons Learned:
Research is hard work, but I prefer doing research to taking classes because it gives me creativity, a sense of purpose, and flexibility.

Fall 2016

Accomplished:
Started my Master's in Computer Science and completed the following courses: Algorithms Design and Analysis, Advanced Operating Systems, and Machine Learning. Also started working with Dr. Filippo Menczer on Fact Checking, later publishing a paper in WSDM Cup 2017.
Lessons Learned:
Graduate school is hard work, but surrounding oneself with what one values most can make the time more pleasant. For me this was doing research and outdoor activities such as running, climbing, tennis, basketball, etc...

2014-2016

Accomplished:
Worked as a Consultant, later being promoted to Senior Consultant at Avanade, a technology company. I learned a lot about dev-ops, the application development lifecycle, team leadership, Web Development, and Big Data here.
Lessons Learned:
Working in industry is good money, but if one wants to make a difference in the world then a career in research, nonprofit, or entrepreneurship would be a better fit. This is when I chose to pursue a research-oriented Master's degree in Computer Science.

2010-2014

Accomplished:
Graduated from Purdue University with a Bachelor of Science in Mathematics. I also gleaned some neat skills in Computer Science, by attending an REU at University of Michigan and working with professor Sorin Matei on a publication about the Iraq War.
Lessons Learned:
The meaning of giving back through Purdue's EPICS program. NOT to procrastinate. Critical thinking is the key to a bachelor's degree, everything else can be forgotten.
Connect with me via: