Back close

PhD Positions with Teaching Assistance Fellowships

Need for trained research scientists/faculty in the area of AI and Data Science is increasing. All engineering and science disciplines are becoming more and more data driven and demands all faculty to be proficient in emerging AI technologies. Modern industry requires more AI-aware R&D engineers than software developers. The experience you gain from the center will equip you to occupy those positions.

This is an opportunity for you to be part of a research center with very vibrant and transdisciplinary research culture. The center is having faculty from all disciplines with very high proficiency in computing.

Typically, you will be able to finish your PhD in 2.5 to 4 years.

Fellowship amount is Rs. 25,000 per month

 

Eligibility Criteria:

  • The candidate should possess an M.Tech degree in Computer Science/Signal Processing/ Data Science/Artificial Intelligence/Computational Engineering/Electrical and Electronics Engineering/ Electronics and Communication Engineering/Information Technology/Bioinformatics/Biotechnology/Mechanical Engineering/Chemical Engineering
  • Programming in Python/MATLAB, Knowledge in AI, Machine Learning and Deep Learning

 

 

Faculty who is offering PhD positions.

 

Dr. Jyothish Lal G

Assistant Professor, Center for Computational Engineering and Networking (CEN)

  • Amrita Profile:https://www.amrita.edu/faculty/jyothish-lal-g
  • Google Scholar:https://scholar.google.co.in/citations?user=QC0uozEAAAAJ&hl=en
  • E-mail:g_jyothishlal@cb.amrita.edu
  • Research Interests:Speech Signal Processing, Nonlinear Techniques for Time-series Analysis, Machine Learning/Deep Learning for Speech Technology Applications
  • Research problem 1:
      • Emotion Recognition : A complex system approach

    Much of the research in speech signal processing is based on the linear techniques by approximating the speech as produced from a linear system. However, since speech is produced from a nonlinear and complex system, such approximations may not be accurate in extracting the information from the speech signal. Therefore, this research aims to analyse the speech signals, precisely, the emotive speech signals in the Indian context from a complex system perceptive. The analysis is expected to extract complex network features that can be used for training machine learning/deep learning models for improving the accuracy of emotion recognition from continuous speech data. The research also requires data speech data collections for multiple Indian languages such as Tamil, Malayalam, Telugu etc.

  • Research problem 2:
      • Speech recognition: Considering the variabilities of age, accent, and disability

    Speech recognition is the conversion of the speech signal to its corresponding text. Conventional approaches to find solution to this problem use features derived based on linear theory of speech production. System based on these features may not always give promising results for the wide variations occurring due to difference in age, accent, and disability of the person. This research aims to address these variations in the speech data and provide a data driven solutions to it. The research also includes data collections for the aforesaid speech variations.

  • Research problem 3:
      • Speech pathology detection: Investigating the significance of nonlinear features

    Traditional methods in this direction of research aims to estimate the glottal wave parameters to detect the disorder from the speech signal. Most often, they employ linear prediction (LP) analysis to estimate the excitation signal or models estimated from the LP residual. Nevertheless, the features estimated using these methods tends to be erroneous and therefore may not provide consistent results. At this onset, there is a scope for investigating on new methods for glottal flow estimate and estimation of accurate glottal parameters. Tools and measures from nonlinear dynamics can therefore be explored for accurate parameterization of glottal waveforms and detection of voice pathology. A database of isolated vowels and consonant-vowels in the Indian context has to collected from subjects for the study.

 

Dr. Neethu Mohan

Assistant Professor, Center for Computational Engineering and Networking (CEN)

  • Amrita Profile:https://www.amrita.edu/faculty/m-neethu
  • Google Scholar:https://scholar.google.co.in/citations?user=B6zK9XYAAAAJ&hl=en&oi=ao
  • E-mail:m_neethu@cb.amrita.edu
  • Research Interests:Signal Processing, Artificial Intelligence, Machine Learning, Computer Vision, Biomedical Image Processing
  • Research problem 1:
      • Development of Adaptive and Intelligent Methods for Power Quality Monitoring

    The objective is to develop adaptive and intelligent methods for power quality monitoring in emerging power grids. The transition from the conventional grid to the modern, intelligent grid known as smart grid is influenced by several economic, social and technical factors, that brought different stages of automation, transformation and evolution in the grid. Smart grid combines myriad of technologies like information, communication, and power system engineering with computational intelligence. Thus, more sophisticated techniques are required to monitor, diagnose, control, and protect the grid operation and its behaviour. Considering these aspects, the current work focuses on developing adaptive and intelligent methodologies for power quality monitoring. The methodology will comprise of combining methods from signal processing, deep learning, artificial intelligence etc. Publication of results will be an integral part of the PhD program

  • Research problem 2:
      • Artificial Intelligence for Biomedical Signal Processing

    Biomedical signals convey information about the physiological processes and underlying abnormalities of human body. Processing of the biomedical signals collected using multiple sensors provides vital information about the human health. However, analysis of data using traditional techniques are time consuming and ineffective due to the availability of vast amount of recorded data and limited resources. Thus, more advanced automated AI assisted systems are required to reveal significant information related to the human health. Motivated by these facts, this research work focuses on the development of AI assisted systems for biomedical signal processing. The methodology will comprise of combining methods from signal processing, deep learning, artificial intelligence etc. Publication of results will be an integral part of the PhD program.

 

Dr. Sachin Kumar S

Assistant Professor, Center for Computational Engineering and Networking (CEN)

  • Amrita Profile:https://www.amrita.edu/faculty/sachin-kumar-s
  • Google Scholar:https://scholar.google.co.in/citations?user=-ztk7jMAAAAJ&hl=en
  • E-mail:s_sachinkumar@cb.amrita.edu
  • Research Interests:Understanding Cognitive Intelligence – theoretical aspects and applications, Artificial Intelligence/Traditional Machine Learning/Deep Learning, Natural Language Processing, Optimization based data analysis (Signal and Image)
  • Research problem 1:
      • Emotion AI, detecting Emotions from text, physiological signals etc

    The main aim of the PhD work will be to develop emotion-sensing machine learning models. Emotion AI refers to the detection and interpretation of emotions from the data sources such as text, voice/audio, video etc. This research area is gaining lot of attention as it finds application related to advertisement, automotive, health monitoring systems etc. The methodology will comprise of combining methods from signal processing, natural language processing techniques, deep learning etc. Publications of research findings will be an integral part of the PhD program.

  • Research problem 2:
      • Deep Visual-Semantic Embedding for Improved Word Representation in Indian Languages

    The objective is to develop a deep learning based visual-semantic embedding for improved word representation in Indian language. The visual world is getting populated every day under several category of objects. State-of-the-art systems tries to solve the recognition tasks by considering labelled data over a defined class. This approach leads to the development of several categories of data over defined class, modelling visual-recognition system based on it. However, scaling such system by including several other categories is still a research problem. As the model is trained by considering the fact that each category of objects are disconnected, having its on behaviours or characteristics. This creates a system which do not extends or transfer the knowledge learned in the context of unseen words. Considering the importance of natural visual-space instead of artificial disconnected categories, the current work proposes a visual recognition model incorporating the text-semantic space, particularly focusing on the Indian Language context. The methodology will comprise of combining methods from signal processing, natural language processing techniques, deep learning etc. Publications of research findings will be an integral part of the PhD program.

 

Dr. Nimal Madhu M

Assistant Professor (Sr. Gr), Center for Computational Engineering and Networking (CEN)

  • Amrita Profile:
  • Google Scholar:https://scholar.google.com/citations?user=LdwjpcIAAAAJ&hl=en
  • E-mail:m_nimalmadhu@cb.amrita.edu
  • Research Interests:AI and Machine Learning Applications to Power Systems; Renewable Generation & Demand Forecasting; Deregulation and Power Systems Economics; Risk Analysis and Portfolio Optimization; Microgrids and Active Distribution Networks; Second-life Battery for Microgrid Storage; Power System Modelling, Analysis and Optimization; Hybrid & Renewable Energy Systems Design & Optimization; Available Transfer Capability.
  • Research problem 1:
      • Wide Area Monitoring and Protection with AI for Power System Stability

    With the use of PMU units for wide area monitoring, compiling and analysis of the location and time-stamped data measure by these devices are providing avenues for not just the application of AI, but for further exploration too. This proposed research endeavours to assess the stability of high and low voltage (microgrid) systems with high penetration of renewable sources, all the while including the stochastic nature of the load as well. The state transitions of the system will be investigated with feature extracting procedures and shall be used for proactive protection of power systems. It is expected to produce two models, suitable for aiding the system stability as mentioned before, as well as for the constrained generation scheduling while accounting for the uncertainty in various components.

  • Research problem 2:
      • Markov Chain-based Network Modelling for Transient and Steady State Simulations

    The application of Markov chain/ Markov models for modelling the operating regime of a microgrid (low or medium voltage system) is the core of this research. The study needs to be carried out breaking down the different components of the network that vary irrespective of the other influencing factors. Hence a layered structural model needs to be built such that the complete dynamics of the system is mapped. The LV system models obtained as outcomes of this should be considerate to the islanded mode of operation of the system as well. The satisfactory modelling will pave way for exploring the working of the model with HV systems and could lead to collaboration with software developers.

 

Dr. Premjith B

Assistant Professor, Center for Computational Engineering and Networking (CEN)

  • Amrita Profile:https://www.amrita.edu/faculty/b-premjith
  • Google Scholar:https://scholar.google.co.in/citations?user=B4J7II0AAAAJ&hl=en
  • E-mail:b_premjith@cb.amrita.edu
  • Research Interests:Natural Language Processing, Computational Linguistics, Cybersecurity, Bioinformatics, Machine Learning, Deep Learning
  • Research problem 1:
      • Natural Language Processing for studying the patterns in Biochemical entities

    Recent advancements in Natural Language Processing (NLP) accelerated the application of NLP to generate representations for words, which can explain the semantics of the words effectively. Biochemical entities such as proteins and chemicals are represented using a sequence of characters and can be considered as words in a language. Text-based representations of biochemical entities using NLP can describe the knowledge about their language. NLP algorithms for generating text representations can elucidate the hidden knowledge in the proteins and chemicals, which help to understand their properties, structure and dynamics. The purpose of this research is to explore the NLP techniques to understand the underlying patterns (properties, structure and dynamics ) in proteins and chemicals.

  • Research problem 2:
      • Natural Language Processing for CyberSecurity

    Log files stores all the activities happening on a computer. These files are text files that contain information such as running applications, system services, errors in the system, and the communication between kernel and processes. Due to this reason, log files are said to contain security information about a system. Analysing the system log files help to identify the attacks and unauthorized access to the system. This research aims to analyse the system log files using Natural Language Processing (NLP) methodologies and thereby prevent attacks to a system.

 

Dr. Harishchander Anandaram

Assistant Professor, Center for Computational Engineering and Networking (CEN)

  • Amrita Profile:
  • Google Scholar:https://scholar.google.com/citations?user=mokZJ14AAAAJ
  • E-mail:a_harishchander@cb.amrita.edu
  • Research Interests:Bioinformatics, Computational Systems Biology, Functional Genomics and Metabolic Engineering
  • Research problem 1:
      • Regulatory Analysis of Circular and Long non-coding RNAs in Pathogenesis of Cancer and Auto Immune Disorders

    Regulatory Networks play a vital role in the process of Gene Regulation in understanding differential expression. Gene Regulation is associated with the molecular functions and various biological processes of various organisms. The regulatory dynamics involve multiple interactions between gene, RNAs and Transcription Factors. To decipher the complexity and reveal potential targets regulatory analysis of multiple networks by statistical and experimental methods will initiate the discovery of therapeutic biomarkers.

     

  • Research problem 2:
      • Biologics to tackle COVID

    SARS-Cov-2 is a major threat to humans and rapid polymorphisms cerate drug resistance to small ligands, in order to initiate the process of alternative therapy, biologics has an important application because biologics have a greater bioavailability than small ligands. Among various biologics, therapeutic water soluble peptides will be a good source to tackle the multiple polymorphisms observed in SARS-Cov-2 because the basic to initiate molecular drug resistance is compatible with bio active water soluble peptides.

  • Research problem 3:
      • Compacting hypothetical proteins in Banana Virus and Pseudomonas Aeruginosa

    Hypothetical Proteins play a major role in pathology. Though the basic sequencing data is available for hypothetical proteins their structures were not determined experimentally, molecular modeling and simulation studies can reveal the nearest experimental structure and activity relationships in physiological and pathological conditions can be predicted based on the modeled structure and activity relationships. Simulations in larger time scales can also predict the bioavailability and activation. The outcome of this study will result in a potential synthetic inhibitor to tackle proteins associated with the harmful viruses.

 

Dr. ‪Sanjanasri J.P

Assistant Professor, Center for Computational Engineering and Networking (CEN)

  • Amrita Profile:
  • Google Scholar:https://scholar.google.com/citations?user=6HbWdmcAAAAJ&hl=en
  • E-mail:jp_sanjanasri@cb.amrita.edu
  • Research Interests:Natural Language Processing, Deep Learning, Machine Learning
  • Research problem 1:
      • Ai-based Multilingual Human-Machine Communication

    Natural Language Processing, Knowledge Representation, Dialogue Management Communication between a man and a computer program is one of the long-term goals of the NLP field. Current systems related to chatbot-style communication can respond satisfactorily to a broad range of questions. However, the background of such communications is kept mainly on the lexical level—recent results in the task of open domain question answering promise to bring adequate knowledge to such dialogues. The thesis combines the two approaches (general discussion robots and question answering systems) in a new practice concentrating on a multilingual environment. In the evaluation part, the thesis needs to offer new results (also) for languages other than the mainstream ones to prove their applicability to a broad spectrum of languages.

     

 

Dr. Navaneeth Haridasan

Assistant Professor, Center for Computational Engineering and Networking (CEN)

  • Amrita Profile:
  • Google Scholar:https://scholar.google.com/citations?user=67HEC3MAAAAJ&hl=en
  • E-mail:h_navaneeth@cb.amrita.edu
  • Research Interests:Computational Mechanics, Computational biophysics, Nanofluidics and Molecular dynamics
  • Research problem 1:
      • Low cost Nanopore sensors for particle detection

    Nanopore sensors are new generation sensors intended for detection of biomolecules such as nucleic acids and amino acids. Solid-state nanopore sensors have the potential to become superior sensing devices in the future, being versatile in terms of size, shape, strength and functionalization. A low-cost fabrication technique to produce nanopore sensors is planned. The nanopore sensor manufacturing will be automated in order to produce the device at a larger scale.

     

  • Research problem 2:
      • AI assisted biomolecular simulations

    Molecular dynamics simulations have now become an inevitable computational tool for exploring many biological processes such drug discovery, sequencing and sensing of biological entities. Combining AI with molecular simulations, the computational method becomes extremely useful in exploring phenomena unfeasible to study using experimental methods. Integrating AI into protein molecular dynamics simulation for the purpose of sequencing or/and property estimation is currently planned.

 

 

Interested candidates can send their resume

Apply Online

  • Add File or drop files here
    Upto 500kb | doc, docx & PDF format only

  • To confirm your request , please check
    the box to let us know you are human
Admissions Apply Now