The admission process for the Mid-year 2022 cycle to the Department of Computational and Data Sciences (CDS) comprises two phases. Phase I will be an **online aptitude test** on the HackerEarth platform, which serves as the first part of the evaluation and Phase II will be an **in-person oral interview** for all candidates who qualified in the aptitude test conducted in Phase I.

The aptitude test is scheduled on **12 Nov 2022 (Saturday) at 11am**. You will receive a separate email from the CDS department for your test and instructions with the test link, etc. This online aptitude test is of **45 minutes** duration. This will have simple multiple choice and programming questions as per the syllabus given below.

The Phase II ** in-person oral interview** is only for students who attempt and qualify in the online aptitude test conducted in Phase I. The day after the aptitude test, an email with the results will be sent to you. In case you qualify, your interview date and session will be as listed in the call letter from IISc Admissions Office. The syllabus for the interview (similar to the aptitude test, with lab-specific readings) is also given in detail below.

The shortlisted candidates will also be sent an email from the CDS department to fill out an **online student information Google form**. This must be completed and submitted in order to participate in the oral interview. As part of this, candidates should choose up to three labs for which they will be considered. These labs are described below.

**Online Aptitude Test**

The Phase I online objective and programming test will be conducted using the “HackerEarth” online platform. Candidates will need a **desktop/laptop with an internet connection, Chrome, Firefox or Safari browsers, and a webcam**. Candidates will be sent an email with a link to the online test on Friday, Nov 11, the day before the test. The test will be enabled at the start of your session and will last for **45 mins.** It must be completed in a single sitting. You will need to create a **free HackerEarth account** to take the test. You MUST use the **same email address** as the one used for your IISc Admissions Application. Please do this ahead of time. That email will also have another link to let you take a **practice test**, to test your browser and webcam, and get familiar with HackerEarth. *These practice questions are not from the test syllabus.* Candidates may take the practice test as many times as they want. Candidates should get familiar with the programming IDE, compilation, and testing interface of HackerEarth to ensure that the programming questions are attempted properly.

The online aptitude test will have 12 questions. There will be 10 multiple choice questions (5 marks each) and 2 programming questions (25 marks each). The **topics **for the objective test are:

- Combinatorics, Linear Algebra/Matrices, Probability and Statistics, Differential Equations, Plotting, Data Structures and Algorithms.

The programming languages that you can use are **C, C++** and **Java**. The HackerEarth interface will allow you to type the program, compile it and run test cases, all within the browser. We will provide a basic code template for reading and writing the inputs and outputs for each problem. The candidates should focus on the actual code logic and brush up on concepts such as if/then/else conditions, for/while loops and arrays/data structures.

**Sample questions** from previous year tests are available here.

**In-Person Oral Interview**

Students who qualify in the online aptitude test (Phase I of the interview) will attend an in-person oral interview at the CDS Department, IISc (Phase II of the interview). In the oral interview, you will be questioned on the following basic topics, and on advanced topics that are based on your lab preferences:

** Basic Topics**: Linear Algebra; Probability & Statistics; Programming, Data Structures, Algorithms; Ordinary Differential Equations; Discrete Mathematics. Final year undergraduate level preparation is required.

** Advanced Topics: **You will be questioned on the topics related to labs selected in the student information Google form. The topics for each lab and the prior training expected are listed below. Candidates choosing a lab must be prepared on

*at least one*of the lab-related interview topics.

Candidates should prepare well on their fundamentals, come with a focus and knowledge of the research areas they would like to target in their PhD, and have clarity on the preferred lab(s) and the kind of work done in them. Review the research topics and the papers from the labs you are interested in.

**Orientation to the Labs**

**Please review these videos to get familiar with the labs taking students during this mid-year admissions.**

- DREAM https://www.youtube.com/watch?v=rBLKTxPpaL4
- VAL https://www.youtube.com/watch?v=W1SPpeP6sVI
- MARS https://www.youtube.com/watch?v=6yhkcWWu-b4
- VCL https://www.youtube.com/watch?v=jic4xdRsZAg
- NATL https://www.youtube.com/watch?v=fH8R5qIS9_w

**Labs Taking Students in Mid-year Admissions, 2022**

**1. Numerical Algorithms and Tensor Learning Lab (NATL Lab)**

**Faculty**: Ratikanta Behera

**Website**: http://cds.iisc.ac.in/faculty/ratikanta

**Lab: **http://cds.iisc.ac.in/faculty/ratikanta/lab.html

**Lab Description: **Our research interests are in the field of scientific computing with particular emphasis on the wavelets methods, numerical multilinear algebra, and neural networks. In numerical multilinear algebra, we develop novel HPC-driven algorithms and theories for solving important existing and emerging multidimensional problems. Specifically, interested in developing fast tensor algorithms for solutions to multilinear systems, nonlinear optimization problems, low-rank approximation, generalized inverses of tensors, and solutions to partial differential equations in high dimensions problems. In wavelet methods, we propose HPC-centric adaptive wavelet methods and algorithms for solving partial differential equations and integral equations, data compression, signal recognition, and signal & image processing. Further, in neural networks, we design adaptive recurrent neural networks to solve time-varying problems. In particular, we aim to prove theoretically and numerically the behavior of adaptive recurrent neural networks under various activation functions.

**Interview Topic: **Numerical Analysis, Linear Algebra, Differential equations, and basic programming.

**Student background: **Students must have a Bachelors’ or Masters’ degree in Mathematics, Physics, Computer Science, Earth Science, and Mechanical, Information Technology, Electrical, Electronics and/or Communications. Basic Programming skills (in any programming language) are required.

**2. DREAM:Lab (Distributed systems Research on Emerging Applications & Machines)**

**Faculty: **Yogesh Simmhan

**Website: **http://cds.iisc.ac.in/faculty/simmhan

**Lab Description: **Focuses on distributed systems, cloud computing, graph algorithms and scalable software platforms. Research topics on software platforms for Cloud and Edge computing and Internet of Things (IoT); Scalable platforms for distributed and federated deep learning; Distributed software platforms and algorithms for drones/UAVs and video analytics; Storage, query and management of Big Data; and Distributed and Streaming Graphs Algorithms, and platforms for Graph Neural Network.

For the 2022 Admissions cycle, the DREAM:Lab is looking for 3-4 students interested in:

- Federated Deep learning platforms using GPU-accelerated edge and cloud computing
- Serverless platforms for multi-cloud and edge computing
- NoSQL and Big Data platforms for spatio-temporal and video data
- Distributed analytics and coordination algorithms for drone fleets
- Machine learning for optimizing cloud and distributed applications
- Incremental algorithms for large graphs and distributed platforms for graph neural networks

*Students must have a strong aptitude in systems software, algorithms and programming.*

**Interview topics: **Operating Systems (or) Graph Algorithms (or) Distributed Systems.

**Recommended Reading:**

- Operating System Concepts by Silberschatz, Gavin and Gagne
- Introduction to Algorithms: A Creative Approach, Chapter 7: Graph Algorithms, by Udi Manber
- Distributed Systems: Concepts and Design by Coulouris, et al.
- Pregel: A System for Large-Scale Graph Processing, Malewicz, et al.,
*SIGMOD*2010

**Prior training: **Students must have a Bachelors’ or Masters’ degree in Computer Science, Information Technology, Electrical, Electronics and/or Communications. Strong programming, algorithms and systems skills are required.

**3. Middleware And Runtime Systems (MARS) Lab**

**Faculty: **Sathish Vadhiyar

**Website: **http://cds.iisc.ac.in/faculty/vss

**Lab: **http://mars.cds.iisc.ac.in

**Lab Description: **High performance computing (HPC), Parallel computing – middleware, system software, algorithms and applications on large-scale parallel computers and GPUs.

Some of the recent research topics in the MARS lab are the following.

- Identifying scalability bottlenecks, and providing high performance and scalable solutions for parallel scientific applications.
- High performance parallel I/O solutions for scientific and ML/DL applications.
- Harnessing the power of both CPUs and GPUs in a hybrid manner in modern-day parallel systems.
- Novel models of parallelism for Machine Learning (ML)/Deep Learning (DL) applications.
- Communication-minimization strategies for Exascale applications including development of asynchronous methods, one-sided communications and approximate computing.
- Fault tolerance frameworks for parallel scientific and ML/DL applications.

**Interview topics:**

- Operating Systems including pthreads, filesystems.
- Computer architecture including multi-core CPUs and modern GPUs, differences between shared memory and distributed memory parallel architectures, different network topologies (e.g., ring, mesh etc.)
- Graph Algorithms (or) MPI parallel programming interface (Google for “MPI Complete reference” and read Introduction, Point-to-Point and Collective Communications chapters.

**Prior training: **Students must have a Bachelors’ or Masters’ degree in Computer Science, Information Technology, Electrical, Electronics and/or Communications. Strong programming, algorithms and systems skills are required.

**4. Video Analytics Lab (VAL)**

**Faculty: **R. Venkatesh Babu

**Website: **http://cds.iisc.ac.in/faculty/venky/

**Lab: **http://val.cds.iisc.ac.in

**Lab Description: **Deep Learning for Computer Vision, Representation Learning, Domain Adaptation, Adversarial Learning, Self-supervised and unsupervised learning, Object Detection, Learning on long-tail data, Generative modeling, 3D reconstruction

**Interview topics: **Linear Algebra, Probability and Machine Learning basics.

**Prior training: **Students must have a Bachelors’ or Masters’ degree in Computer Science, Information Technology, Electrical, Electronics and/or Communications or in any other closely related areas.

**5. Visual Computing Lab (VCL)**

**Faculty: **Anirban Chakraborty

**Website: **https://anirbanchakraborty.github.io/

**Lab: **http://visual-computing.in/

**Lab Description: **At VCL, we are interested in developing novel computer vision and machine learning algorithms to solve visual analytics problems arising from real-world applications. The majority of our ongoing research projects can be categorized into one of the following three areas – 1. Data-efficient deep learning (zero-shot/few shot learning etc.), 2. Learning across modalities/domains (e.g., text-based image retrieval, sketch-guided localization, visual question answering, domain adaptation etc.) and 3. Video surveillance (e.g., person-reidentification, anomaly detection, human gait and pose analysis etc.)

**Interview topics: **Linear Algebra, Probability and Machine Learning basics.

**Prior training: **Students must have a Bachelors’ or Masters’ degree in Computer Science, Information Technology, Electrical, Electronics and/or Communications or in any other closely related areas. Strong programming and algorithms skills are necessary. Prior experience in image/video analysis and/or machine learning would be a plus.