Somayyeh Koohi's Homepage/Students

Department of Computer Engineering

Director

Dr. Somayyeh Koohi

PhD Students

Mahmood Kalemati (An efficient solution for drug target interaction and binding affinity prediction using deep learning methods )

Kosar Yousefi-nezhad

Sahebeh Nabavi

Fatemeh Moradi

MSc Students

ALireza Belal (Drug-target binding affinity prediction based on machine learning algorithms and features extraction using neural networks )

Aref Shahbakhsh (Prediction of mRNA subcellular localization using deep learning and Investigating the possibility of optical implementation )

Hadis Ahmadian (Improvement of deep learning based methods for Protein Structure Prediction)

Mohammad-Amin Fereidoon (Providing a tool for predicting the tertiary structure of proteins using neural networks and coding the sequence of proteins based on GPU )

Tala Kiani (Redesign of the parallelized Kraken algorithm with the aim of achieving memory efficiency for data classification )

Azam Gholami (Genome Sequence matching based on Hyper-Dimensional Computing with computing in memory capability )

Farhan Saranavnd (A machine learning-based method for antibody paratope prediction using ensemble learning and evolutionary-based features )

Hadi Jasouri (Proposing an optimized hyper-dimensional computing platform customized for biosignals using in-memory processing )

Sara Kesshavarz (Read mapping from scRNA-Seq data to a reference sequence with optical implementation capability )

Vahid Moghimi (Improving interpretability of protein function predicting model from protein sequence using deep learning methods with optical implementation capability )

Ali Sedaghi

Haniyeh Molaei

AmirHossein Alishahi

Mahdi Hassanzadeh

Zahra Chizari

Graduated PhD Students

Saeedeh Akbari (Biological Sequences Comparison Based on Optical Processing)

Hoda Sadeghzadeh (Design of optical convolutional nueral network for image classification)

Melika Tinati (Advisor)

Elham Khani (Co-Superviser)

Graduated MSc Students

Ali Dehghan Nayrei (Protein interaction prediction based on deep neural networks through efficient FPGA and GPU implementation )

Maryam Gheysari (Drug-target Interaction Prediction through Learning Methods for SARS-COV2 Based on Sequence and Structural Data)

Aida Ebrahimi (Designing an optical processing unit for non-Linear operations in deep neural networks)

Azam Abdosalehi (A Fast Alignment-free protein Comparison approach based on FPGA Implementation)

Farzin Mohammadi (DNA motif finding using edit-distance approach)

AmirHossein Mohammadi (Developing a deep neural network for bio-sequence classification capable of optical computing)

Fatemeh Tabatabaee (prediction of DNA/RNA sequence binding site to protein with the ability to implement on GPU)

Mojtaba Zamani (Drug target binding affinity prediction using a deep generative model based on molecular and biological sequences)

Pouria Laghaee (Provide tools for clustering single cell data with low computational complexity)

Reyhaneh Ahmadi (Design of free-space optical spiking neural network)

Saeed Darvishi (MHC-Peptide binding prediction using a deep Learning method with efficient GPU implementation approach)

Roshanak Karimi (Reliable Adaptive Wavelength Modulation for Optical Networks-on-Chip )

Negar Rezaei (Adopting Dynamic Topology for Energy Management in Optical Interconnection Networks in Data Centers )

Hossein Babashah (Design and analysis of DNA string matching by optical parallel processing )

Ehsan Melaki (Design and implementation of DNA pattern recognition algorithm utilizing optical coding methods )

Saeedeh Akbari (Designing a Fault tolerant Optical Interconnection Network for Data centers )

Ali Nezhadi

Ali-Reza Abolhasani (Fault Rate Modeling in terms of Power Consumption and Thermal Variation in Optical Networks-on-Chip)

Alborz Derakhshan-far (Optical Interconnection Network for Data centers)

Hamid-Reza Erfani (Thermal-Aware Routing Algorithm in Fault-Tolerant Optical Networks-on-Chip )

Mahdieh Movahederad (All-Optical Scalable Multi-Stage Interconnection Network for Data Centers )

Reza kalhor (DNA Classification Using Optical Processing based on Machine Learning)

Azhar Abbas (Energy Aware Routing Algorithm with SDN in Data Center Networks)

Research Projects

Bio-Photonics

Bioinformatics is an interdisciplinary field that expands methods and tools for biological data comprehension, and benefits from several sciences, such as computer science, statistics, mathematics, and engineering to expound and analyze biological data . As an important operation in bioinformatics, sequence alignment arranges the sequences of DNA to identify regions of similarity and differences between the sequences to detect the genetic disease. Each DNA sequence is represented with a string of A/G/T/C characters; each is called a nucleotide, and composed of either of adenine (A), guanine (G), thymine (T), or cytosine (C). As the main purpose of sequence alignment algorithm, permanent change in the nucleotide sequence, denoted as mutations, should be located. Mutations are classified either as Small-scale mutations or as Large-scale mutations. A small scale mutation which affect a small gene in one or a few nucleotides causes a nucleotide base substitution, insertion, or deletion of the genetic material, DNA or RNA . The mutations, which happen randomly, are being recognized as particularly important in many genetic disorders. Therefore, detecting and locating point mutations may be known as the most important challenge for detecting genetic diseases.

The rapid growth of DNA sequence data in biology databases has necessitated efficient collating, organizing, identifying, retrieving, and searching the sequences as DNA representation. It’s worse when a large amount of data should be processed in real-time. However, considering big-data processing, traditional electronic computers suffer from many limitations including high power consumption, heat generation, high delay and slow response .

To overcome electrical computer limitations, optical computation has been proposed . Capability of parallel processing in optical computers motivates us to optically implement a global and local sequence alignment method. For this purpose, optical correlation benefits from the potential of high speed processing can be employed to perform rapid global alignment for detecting high similarity among various sequences.

Optical Supercomputers

The growing technologies have increased the demand of high performance computing. According to G. Moore’s low, number of transistor counts to be integrated per unit area in devices will almost double in one and half year. To achieve high speed computation, high packaging density in the logic circuits is required which results in more heat dissipation. The conventional computing is found unable to deal with low power, high compaction and heat dissipation issues of the current computing environment.

Optical Computing is computation with photon as opposed to conventional electron based computation. Unmatched high speed and zero mass of photon have attracted the researchers towards the optical realization of reversible logic gates using Semiconductor Optical Amplifier (SOA) based Mach Zehnder Interferometer (MZI) switches. In order to build the optical computer, all-optical flexible signal processing devices are needed. Optical logic gates are considered as key elements in real time optical processing and communication systems which perform the necessary functions at the nodes of network such as data encoding and decoding, pattern matching, recognition and various switching operations.

Emerging Technologies for on-Chip Communication Architectures

The increasing interconnection bandwidth demands for chip multiprocessors (CMP) cannot be simply satisfied by the reduction of the transistor feature sizes and raising of the chip operation frequency. This problem stems from many limitations associated with electrical interconnects, such as latency, impedance, bandwidth, and power consumption.

Optics provides low power dissipation that remains independent of capacity and distance, as well as wavelength parallelism, ultra-high throughput, and minimal access latencies. Importance of power dissipation in on-chip communication architectures, along with power reduction capability of on-chip optical interconnects, offers Optical Network-on-Chip (ONoC) as a new technology solution which can introduce on-chip interconnection architecture with high transmission capacity, low power consumption, and low latency. The aim of this project is to investigate innovative interconnect technologies such as photonic waveguide based and 3D hybrid on-chip communication architectures.

Optical Interconnection Network for Data Centers

Data centers are experiencing an exponential increase in the amount of network traffic that they have to sustain due to cloud computing and several emerging web applications. To face this network load, large data centers are required with thousands of servers interconnected with high bandwidth switches. Current data center networks, based on electronic packet switches, consume excessive power to handle the increased communication bandwidth of emerging applications. Optical interconnects have gained attention recently as a promising solution offering high throughput, low latency and reduced energy consumption compared to current networks based on commodity switches.

This project discusses the need for a shift of the data centres networks to the optical domain to reduce the power consumption and meet the bandwidth requirements. The aim of this project is to present some indicative optical interconnects for high performance data centres that have appeared recently in the research literature.
Modeling of Optical On-Chip Communication Architectures

Many of the modern Systems-on-Chip integrate a high density of heterogeneous components such as different processors, a wide range of hardware components, as well as complex interconnects that use different communication protocols. On-chip physical interconnections represent a limiting factor for the performance and energy consumption. Currently, the optical interconnects integrated on chip are a viable alternative for on chip interconnects. However, the access to physical prototyping of these interconnects is a major challenge because this systems require very recent technologies, still difficult to access. Thus, their high-level modeling and validation are mandatory. The aim of this project is to propose a modeling approach of the passive integrated photonic routing structures considering physical-level delay and power issues.

Last Updated : April 2024