image

Associate Professor
EE604, Department of Electrical Engineering
Indian Institute of Technology Hyderabad
Email: lakshminatarajan [at] ee.iith.ac.in

Research Interests

Coding techniques, information theory, wireless communications.

Recent Work

  1. Our work accepted for presentation at ISITA’26 analyzed the performance of the low-complexity successive cancellation decoder for Berman codes and their duals (which includes Reed-Muller codes as a special case). We showed that for certain regimes of code parameters Berman codes and their dual codes attain vanishingly small block error probabilities in all binary-input memoryless symmetric channels irrespective of the channel’s noise level (that is, the Bhattacharyya parameter).

  2. We identified low-complexity decoders for some BiD codes. The decoder uses the minimum-weight codewords of the dual code as parity-check in a redudnant Tanner graph to perform message-passing decoding. Towards this, we obtained numerical bounds on the minimum distance of the dual of BiD codes, and for the dual of second-order BiD codes we have a complete characterization of all the minimum-weight codewords. This work will appear at ISIT’26.

  3. We introduced a new family of algebraic codes of length 3m3^m, called BiD Codes (Berman-intersection-Dual Berman Codes), that can be constructed using a kernel of size 3×33 \times 3. These codes provide minimum distances and block error rates similar to Reed-Muller codes at practical block lengths, while in the asymptotic regime they provide a faster growth of minimum distance. Further, these codes are known to achieve vanishing bit error rates in the erasure channel at all rates less than the channel capacity. This work was present at ITW’25.

  4. We identified good subcodes of second-order Reed-Muller (RM) codes by applying the construction of recursive subproducts (see below) to first-order RM codes. These RM subcodes have rates close to second-order RM codes, but with far fewer minimum-weight codewords. This work was presented at ISIT’25.

  5. Our work on a family of codes, called Recursive Subproduct Codes, appeared at ISIT’24. A summary of this work, including codeword error rate comparsion against Reed-Muller and Polar codes in the AWGN channel, is available in these slides.

  6. Our work on Berman codes (their dual codes and a family of related abelian codes) is available in IEEE Transactions on Information Theory. These codes attain vanishingly small BER in the binary erasure channel at all rates less than capacity. A summary of the work is available in these slides.

Openings

PhD and MTech (RA) Positions

Looking for PhD and MTech (RA) candidates to work on problems related to constructing classical and/or quantum error correcting codes, designing efficient decoders and analyzing their performance.
 
No prior knowledge of coding theory or information theory will be assumed from the PhD candidates. PhD students will be provided a thorough introduction to the topic of research, and will be guided through regular discussions and interactions. Mentorship will be provided for writing papers and academic reports.
 
Educational Background: Bachelor’s/Master’s degree in EE/ECE/CS or related areas, or MSc in Mathematics/Statistics/Physics or equivalent. Candidates must satisfy the eligibility criteria set forth by the Department of Electrical Engineering and IIT Hyderabad.
 
Prerequisites: interest to pursue theoretical research, willingness to learn new mathematical topics, basic programming acumen, good understanding of basics of linear algebra and probabilty theory.
 
Interested candidates can send me an email along with their latest CV and grade card.

Internships

If you are interested in pursuing higher degree (especially, if you are currently in an MTech or MSc program), please write to me for internship opportunities. Please include your CV and the latest grade card in the email communication.

Acknowledgment

The support received from the below organizations is gratefully acknowledged

News