Abhinav Aggarwal

Untitled design

Learn. <Master/>. Innovate.
Visa Research
385 Sherman Avenue,
Palo Alto, CA 94306

Email:  abhiag6891 at cs dot unm dot edu
Personal Blog: abhiag6891
Phone: +1-505-573-4914

About me

I am a Ph.D. student in the Department of Computer Science at the University of New Mexico (UNM). I am currently working with Prof. Jared Saia, who is also my research advisor. I also collaborate with Dr. Varsha Dani and Dr. Thomas P. Hayes for my research. I obtained my M.Tech/B.Tech in Computer Science and Engineering from Indian Institute of Technology, Roorkee (IITR) in 2014 under the supervision of Prof. Padam Kumar

I am a mathematics and theory enthusiast. My research topics of interest are Distributed Computing and Security, Randomized algorithms, Resource Competitive Analysis and Machine Learning. Apart from these, I am open to any mathematically challenging problems, especially ones in the field of Combinatorics and Probability Theory.

When I am not busy with my research, I like to spend my time by solving math puzzles, cooking, and traveling. I am also very fond of singing and creative arts and am looking forward to opportunities for flying lessons to pamper my ever-so-long love for airplanes.

Ongoing Research Projects

Security 

Incentive-Compatible Remote Configuration of Security Gateways. Abhinav Aggarwal, Mahdi Zamani, Mihai Christodorescu. (Visa Research, 2017) :

  • Is it possible to stop the flow of attack packets over a network by remotely configuring the firewall settings of the gateway closest to the source of the attack? 
  • What if both the gateway protecting the source of attack packets colludes with the adversary?
  • What if the victim of the *attack* himself has malicious intentions behind stopping the packets from entering the network?  

GenID: Generating Sybil-Resistant Identities. Abhinav Aggarwal, Loi Luu, Mahnush Mohavedi, Mahdi Zamani. (Visa Research, 2017) :

  • Is it possible to establish identities of nodes in a p2p network with an honest majority but without any trusted PKI?
  • Given n parties out of which t are dishonest, can we design the algorithm such that each honest party outputs a set of IDs that contains the IDs of all honest parties but at most one ID for each dishonest party? 

[NC] Interactive Communication for Large Networks. Abhinav Aggarwal, Varsha Dani, Thomas P. Hayes, Jared Saia. (UNM, 2016) : 

  • Can we make any multiparty asynchronous protocol of an unknown length robust to tampering by an oblivious bit-flipping adversary with unknown budget and unbounded computation power?
  • Is it possible that the overhead is optimal to within a constant?

[NC] Secure One-Way Interactive Communication. Abhinav Aggarwal, Varsha Dani, Thomas P. Hayes, Jared Saia. (UNM, 2015) : 

  • Can Alice send Bob a message of an unknown length over a noisy channel that is subjected to tampering by an oblivious bit-flipping adversary with unknown budget and unbounded computation power?
  • Is it possible to guarantee that the overhead is sublinear in the length of the message?

Machine Learning

Battle of the Machines: Introducing Incentive-Driven Verifiable Learning. Abhinav Aggarwal, Mihai Christodorescu, Mahdi Zamani. (Visa Research, 2017) :

  • Is it possible to train a model and verify it to the provider of the training examples (without revealing the model) when both the trainer and the provider are adversarial? 
  • If the provider pays the trainer for this service and the trainer spends some resources in the computation of the model, how much should these costs be so that the above goal is achieved in a way that is fair to both the players?

Markov Chain Monte Carlo

Strong Spatial Mixing in Bounded Degree GraphsAbhinav Aggarwal, Thomas P. Hayes. (UNM, 2017) :

  • Is it possible to show weak-spatial mixing in graphs of max degree 3 using 4 colors?What about strong spatial mixing?
  • Can this generalize to graphs of max degree Δ using Δ+1 colors?

[NC] – Nearly complete.

 

Publications

NOTE: In theoretical computer science, it is standard to list author names in alphabetical order. The citations that do not follow this convention, and instead list authors in order of contribution, are marked with the (*) symbol.

Conference Papers

  • Malik S, Aggarwal A, Sardana A. Novel authentication system using visual cryptography. In Information and Communication Technologies (WICT), 2011 World Congress on 2011 Dec 11 (pp. 1181-1186). IEEE.

Preprints

  • Abhinav Aggarwal, Varsha Dani, Thomas P. Hayes, Jared Saia: Interactive Communication for Large Networks. [arxiv] (This version is old. Will be updated soon.)
  • Abhinav Aggarwal, Varsha Dani, Thomas P. Hayes, Jared Saia: Secure one-way interactive communication. [arxiv]

Talks

  • Competing for bits: A resource competitive framework for multi-party interactive communication. Visa research, June 2017. [slides]
  • Evil Ironman and the Talkative Men of NYC. UNM Stem Research Showcase, Feb. 2017. [slides]
  • Secure Multiparty Interactive Communication with Unknown Noise Rate. Mid-South Theory Day, LSU, Baton Rouge, Louisiana, Dec. 2016. [slides]
  • A Short Introduction to Algebraic Manipulation Detection Codes. UNM CS Theory Seminar, Sept. 2016. [slides]
  • Secure One-Way Interactive Communication. UNM CS Student Conference, April 2016. [slides]

 

Some Interesting Past Projects

Internship Projects

  • Provisioning an Isolation Manager and Studying the Feasibility of Using Linux Control Groups for User-Level Isolation in Search Queries. Abhinav Aggarwal, Rahul Bansal, Chris Bond. (Google Inc., Mountain View, California, USA, 2016)
  • Analyzing Test Frameworks and Dependencies for Porting the PDF-API test code. Abhinav Aggarwal, Ravindra Sushma Bhartiya. (Microsoft India Development Centre, Hyderabad, India, 2013)
  • Optimizing VHD (Virtual Hard Disk) Comparison Tool for Dynamic VHDs. Abhinav Aggarwal, Priyank Gaharwar, Vinod Kancharla. (Microsoft India Development Centre, Hyderabad, India, 2012)
  • An Arduino Based Hardware Prototype for Audio Level Indication (Opencast Matterhorn Project). Abhinav Aggarwal, Jim Greer. (ARIES Lab, Department of Computer Science, University of Saskatchewan, Saskatoon, Canada, 2011)

Multiplicative Weights Update 

Beating the Multiplicative Weights Update Algorithm: A Devil’s advocate approach. Abhinav Aggarwal, José Abel Castellanos Joo, Diksha Gupta, Jared Saia. (UNM, Spring 2017)

Complex Adaptive Systems 

Note: These projects were done with Prof. Stephanie Forrest in Fall 2015 for her course CS523 at UNM.

Evolving Random Sequences Using a Genetic Algorithm. 

  • Is it possible to use a genetic algorithm to evolve a set of binary sequences to being *random* with respect to a given set of tests of randomness and independence?
  • If yes, is the mutation operator more *impactful* here than the crossover operator?

How Infections Spread on Large Networks? 

  • Given a large random graph, what does the SIR dynamics for infection spreading look like? Does it matter how the random graph is generated?
  • What metric over the graph closely predicts the average number of nodes that are affected?

Bacterial Game Dynamics. 

  • Given three different species of bacteria who each have survival rates against each other and some initial distribution of these three kinds in an (n x n)-grid, how does the evolution look like? Can a cellular automaton evolve to a steady state in all cases here?
  • Is it possible to replicate the results on the dynamics of rock-paper-scissors using local dispersal from [this] paper to study the bacterial dynamics better?

Attacking Graph Coloring in Large Graphs Using Genetic Algorithms. 

  • Is it possible to use a genetic algorithm to obtain a perfect coloring for a given graph?
  • How much do the neutral regions and the equilibrium in learning get affected by the diameter and density of the graph? Do things change if the graph is random?

Iterated Maps and Measures of Uncertainty.

Formal Methods

A Study on the Equivalence of Probabilistic Automata. Abhinav Aggarwal, Deepak Kapur. (UNM, Spring 2015)

Comparing Sequential Computer Programs using Truth-Preserving Partial Functions. Abhinav Aggarwal, Padam Kumar. (IITR, Spring 2014)

A Study of Function-induced-orders for termination of recursive programs. Abhinav Aggarwal, Padam Kumar. (IITR, Autumn 2013)

Others

Virtual Differential Storage Based k-Rollback Concurrency Control Algorithm in Distributed Shared Memory Systems. Abhinav Aggarwal, Rupika Srivastava, Kirti Meena, Poonam, Sumit Malik. (IITR, Autumn 2012)

Principles of Flip Theory for the Development of Flip Cipher. Abhinav Aggarwal. (IITR, Spring 2012)

Functional Grammars: Introduction to Implicit Data Structures. Abhinav Aggarwal. (IITR, Spring 2012)

 

Some of my selected Blog Posts

In Progress

Needles and Expectations: 

Is it possible to use linearity of expectations to solve the Buffon’s Needle Problem? What if the sheet on which the needle is thrown contains square grid lines? What about hexagonal grids? 

Not random enough? Populate. Evolve. Done! : 

Is it possible to construct a sufficiently long sequence of bits that is *random* with respect to a given set of tests of randomness and independence? Is this possible for *any* given set T of tests? 

Published

A *bug* in Democracy : 

How dangerous can abstain from voting be for a democracy that’s headed by a greedy leader? Is there a way the leader can take over the wealth of the nation by exploiting the needs of the common people? 

Notorious Coins : 

  • Given a linear arrangement of coins of various denominations and two greedy players who alternate in choosing a coin from either of the two ends, what maximum value can the first player guarantee for himself?

Dr. Blockchain and her Global Clinic:

  • Is it possible to provide a high-quality diagnosis of diseases using blockchains?

Teaching Experience

(TA) Modeling and Simulation (CS-501 IITR, Spring 2014)

(TA) Discrete Mathematics (CS-106 IITR, Autumn 2013)

 

Awards

  • Institute Silver Medal, Indian Institute of Technology, Roorkee, India 2014

Volunteering Work

 

Note : All information/data/files on this page and linked from this page is a property of Abhinav Aggarwal and is copyrighted. Any replication or use must seek permission unless used strictly as a reference for education purposes.