Aditya Vardhan Email & Phone Number

New Delhi, Dl, In

New Delhi, Dl, In

Delhi, India

Delhi, India

United Kingdom

West Delhi, Delhi, India

Noida, Uttar Pradesh, India

My primary objective in SkyteX is to provide vision and navigation capabilities to UAVs.

To predict the motif mutations in gene expression that cause motor neuron degeneration leading to amyotrophic lateral sclerosis (ALS)In my masters’ thesis, I worked with Dr. Kevin Kenna from the bio-informatics department in University Medical Centre – Utrecht (UMCU). My work involved building a deep neural network frameworks which could predict the motif mutations in gene expression that are likely to cause motor neuron degeneration in amyotrophic lateral sclerosis (ALS). Motifs in DNA sequences are single nucleotide subsets of a sequence in which mutations can lead to adverse effects as observed in ALS patients. I hypothesized that the size of mutations to be directly proportional to the known effect of the mutation ranked as per the traditional approaches. Thus, I built two deep learning frameworks: (1) Convolutional neural network and (2) Convolutional neural network + long short term memory networks. Both frameworks when tested on sequences extracted from ALS patients show that most damaging mutations have a very small size of mutation.

Amersfoort

Classifying personality disorders based on scores intrinsic to behavioral traits (such as inhibition, avoidance, approach, anger and anxiety)My second internship involved working with Dr. Barbara Montagne in GGZ central (mental health institute), Amersfoort. My work involved classifying personality disorders based on scores intrinsic to behavioral traits such as inhibition, avoidance, approach, anger and anxiety. I deployed a convolutional neural network (CNN) using behaviour inhibition scores (BIS), behaviour activation system (BAS) and state trait anxiety inventory (STAI) to classify personality disorders and healthy controls with an accuracy of approximately 90%. We also see that these scales can predict higher morph levels (40% and 50% - levels where the presented emotion stimulus is prominent) of the emotion recognition task.

Utrehcht

‘Information matching the content of visual working memory is prioritized for conscious access (Gayet & Paffen., 2013)’In my first internship with Dr. Chris Paffen and his post-doctoral student Yun Ding, I got an opportunity to conduct meta-analysis and replicate the study findings conducted on the paper- ‘Information matching the content of visual working memory is prioritized for conscious access (Gayet & Paffen., 2013)’. My work involved evaluating the time taken by visual working memory to retain information from the presented stimulus.

Master Of Science - Ms, Artificial Intelligence, 7

Bachelor Of Technology, Computer Science