PhD Candidate

Khoury College of
Computer Sciences

Northeastern University

varma.ak@northeastern.edu

I am a PhD candidate working with Dr. Ravi
Sundaram in the Theoretical Computer Science
Group at Northeastern
University. I expect to graduate by *Summer 2024*.

Before joining NEU, I completed my B.Tech. degree (2013-2017) in Information and Communication Technology from DAIICT, Gandhinagar, India (see Bachelor Thesis). I also got a minor in Computational Sciences and still have some lingering interest in areas like High Performance Computing, Complex Networks, and Modelling and Simulation.

- 2023-07-17: An animated version of one of my favorite proofs (without words) of the Pythagoras theorem.
- 2023-07-15: An animation going over the simplest cases of the Master Theorem that I made to learn to use the manim python library.
- 2023-06-13: Homepage for the UG algorithms course I’m teaching this summer is up.
- 2023-04-11: I successfully finished my thesis proposal.
- 2023-03-17: I’m on Mastodon now, so my Twitter will be dormant.
- 2022-10-21: MRS Communications paper accepted.
- 2022-05-23: I presented preliminary
work on
*Machine Learning based Disambiguation of Magnetic Materials*as a poster at MRS 2022 Spring Meeting. - xxxx-xx-xx:
*(Pre-)history.*

I am interested in understanding the theoretical aspects of problems arising in Computer Science. In particular problems related to:

- Machine learning: specifically those related to kernels and random features, overparametrization, convergence and generalization in neural networks.
- Graph algorithms: specifically those dealing with either estimating or leveraging various kinds of graph parameters and utilizing that to solve problems; as well as various hardness results and lower bounds for such problems. I am also interested in how a theoretician needs a heterogeneous toolkit today including techniques generally sequestered into subfields like approximation, randomized, sublinear or fixed parameter tractable algorithms. I am also working on a problem that has to resort to the (mostly heuristic) world of machine learning to solve certain graph problems.

My PhD thesis’ focus is on the heterogeneous toolkit that is needed for graph problems given the prevalence of large data sets as well as problem scenarios that are less structured.

Here are a list of my publications. The symbol indicates alphabetical ordering of authors as is customary in theoretical computer science.

**Toward remote and secure authentication: Disambiguation of magnetic microwire signatures using neural networks**

*Akshar Varma*, Xiaoyu Zhang, Brian Lejeune, Laura Cebada Almagro, Rafael P del Real, Pilar Marin, Ogheneyunume Fitchorova, Laura H Lewis, Ravi Sundaram.

MRS Communications volume 13, pages 16–20 (February 2023). I presented a poster on preliminary work at the MRS 2022 Spring Meeting (May 2022); see slides below.

**Realization Problems on Reachability Sequences**

Matthew Dippel, Ravi Sundaram,*Akshar Varma*

Theoretical Computer Science, Volume 866, Pages 1-13 (April 2021). I presented a preliminary version at the COCOON 2020 conference and it appeared in its proceedings; see slides and video below.

**Attention improves concentration when learning node embeddings**

Matthew Dippel, Adam Kiezun, Tanay Mehta, Ravi Sundaram, Srikanth Thirumalai,*Akshar Varma*.

arXiv preprint (June 2020). In Dec 2018, I gave a short talk (on preliminary work) at ICTS^{fun-fact}during the TBML discussion meeting; see slides and video below.

**Let’s HPC: A web-based platform to aid parallel, distributed and high performance computing education**

Bhaskar Chaudhury,*Akshar Varma*, Yashwant Keswani, Yashodhan Bhatnagar, Samarth Parikh

Journal of Parallel and Distributed Computing, Volume 118, Part 1, Pages 213-232 (August 2018).

**Reachability Problems and Space Bounds**

*Akshar Varma*; under guidance of Prof. Nutan Limaye, IIT Bombay.

Bachelor Thesis.

Previously, I have attended the SPML discussion meeting at ICTS in January 2020; the Workshop on Theory of Deep Learning at IAS in October 2019; the Non-convex optimization and deep learning workshop at MIT in January 2019.

Quite a long time ago, I had attended the NMI workshop on Complexity Theory at IIT Gandhinagar in November 2016 and the Forum for Information Retrieval Evaluation at DA-IICT, Gandhinagar in December 2015.

I was the instructor for CS3000 (UG Algorithms and Data) in 2023 Summer 2 (July-August): Homepage.

I have been a Teaching Assistant (TA) for a variety of Algorithms and Data Structures courses at Northeastern University: CS4800 (UG) in Sep–Dec 2017, CS3000 (UG) in Jan–Apr 2021, CS5800 (graduate) in Sep–Dec of 2019, 2021, 2022, CS5002 (graduate) in Sep–Dec of 2023. I was the head TA in CS5002 (in 2023) as well as all of the CS5800 Algorithms courses. The Algorithms courses contained anywhere between 100 and 300 students in any given semester. Over the numerous offerings I have been responsible for each aspect of the course, including: creating (and deciding a rubric for) problem sets, recitation handouts, exam papers, quizzes, programming assignments, etc.; teaching lectures when the instructor was traveling; running recitation sections; tutoring students from a non-Computer Science background; assisting the instructor in managing workload among the other TAs.

To help with TAing, I made a LaTeX template for easily creating the various handouts seen in such courses. You can find it on my repo of various LaTeX templates; see the course-handout-usage-example.tex file for an example of using the course-handouts-preamble.sty style file.

In the past (Jul–Nov 2016) I have also been a Teaching Assistant for an undergraduate High Performance Computing course offered to juniors. Experiences in teaching this course, and the tools we hacked up to help, lead to the development of a web-based platform to aid HPC education and a journal paper. Details above. ⤴

My Emacs config ^{idea h/t}, on EmacsWiki

- Some of my favorite built-in parts of Emacs include: Ido, Dired, TRAMP, Keyboard Macros, Org+Agenda and most importantly the beautiful self documentation.
- Some packages of note I use from MELPA include: Magit, AUCTex, PDF Tools, Notmuch.
- Packages I use which improves the Emacs (editing) experience without
any significant learning curve. These tend to be among the top
downloaded MELPA packages.
- Plug and play: aggresive-indent, beginend, highlight-numbers, hungry-delete, anzu (just the basic stuff is sufficient), undo-tree, which-key, browse-kill-ring, dired-rainbow, volatile-highlights.
- 10-15 minutes for basic usage: smartparens, expand-region, move-text, wgrep, iedit.

- See also: Emacs
itself is my “favorite Emacs package”

- The
*What’s new*section is new, so what little there is back-filled and is missing stuff and probably incorrect in some ways. - Fun fact: The square and triangles in the ICTS logo provide a visual proof of the Pythagoras theorem. For more fun, here is a page with over a hundred proofs of the Pythogoras theorem.
- Idea from: https://morganastra.me/tools-I-use.html, which does have some nice tools listed in there. I use none of them, although I do use Emacs (instead of Spacemacs), ripgrep/rg in Emacs (an alternative to ag that is available as an Emacs package), Vimium (which is similar to Vimperator). I only list Emacs since I spend most of my time in there and don’t have particularly strong views about any other tool.
- My name, Akshar Varma, in various other scripts that might aid pronunciation. Roughly in decreasing order of confidence. (Note: I don’t actually speak all of these languages.)

Last modified: `October 11, 2023`