Aditya Ahuja
Updates  |  Work Experience  |  Teaching  |  Research  |  Projects

I am second year CS Masters student at UC San Deigo specializing in Computer Vision and Machine Learning. At UCSD, I am affiliated with the Bazhenov Research Group, and advised by Prof. Maxim Bazhenov.

Before that, I completed my undergraduate degree in Computer Science from BITS Pilani, Goa where I was affiliated with APPCAIR (& TCS Research) and supervised by Prof. Ashwin Srinivasan.

I pursued my undergraduate thesis on few-shot instance segmentation models under the supervision of Prof. Hanspeter Pfister at Harvard's Visual Computing Group (VCG).

Before that, I was a research intern at ECMWF, working with Dr. Peter Dueben, and building Deep models to detect anomalies in ECMWF's massive data services. I also spent a summer at, working with the Ad-Experience team on malware detection in web advertisements.

Previously, I was the president of Society for Artificial Intelligence and Deep Learning (SAiDL).

Email  |  Google Scholar  |  Github  |  LinkedIn

profile photo

Check out the CATER Demo and the BlackSwan Demo!

Mar '22

I'll be working with Google Advanced Technology & Projects (ATAP) team this summer as a Computer Vision intern!

Jan '22

Started working at Bazhenov Research Group as a Graduate Research Assistant.

Dec '21

Our paper on Incorporating Domian Knowledge in Neural Networks got accepted at Nature Scientific Reports.

Sep '21

Started my Masters in Computer Science at UC San Diego!

May '21

I've been selected as a Google Summer of Code (GSoC) intern at CERN! I'll be working on normalizing flows.

Jan '21

I've started my undergraduate thesis at the Visual Computing Group (VCG), Harvard mentored by Prof. Hanspeter Pfister!

Nov '20

Check out me and Adithya talking about our work at ECMWF. [Video]

Sep '20

Got a Pre-Placement Offer from for Fall 2021!

Aug '20

Excited to be one of the 150 Indian undergrads selected for the Google Research AI Summer School!

Jul '20

Co-organised the Summer Symposium on AI Research with 3000+ registrations, inviting top AI researchers as speakers.

Jul '20

I'll be a TA for the iXperience summer Data Science program!

Jul '20

Excited to start my research internship at ECMWF as part of their open source summer program - ESoWC!

May '20

Me and Ajay Subramian are mentoring a summer project for the SAiDL-Season-of-Code.

May '20

I'll be interning at (Directi) as part of their Ad-Experience team over the summer.

Jan '20

I'll be working with APPCAIR and TCS Research on Visual Reasoning and Neuro-Symbolic modelling.

Jan '20

I'll be TAing the Machine Learning Course at BITS Pilani, Goa.

May '19

I'll be a mentor for this summers' Machine Learning QSTP course along with Rijul and Saura.

Jan '22 - Current

Bazhenov Research Group - Graduate Research Assistant.

Jun '22 - Sep '22

Google ATAP - Software Engineering Intern.

Jun '21 - Aug '21

Google Summer of Code (GSoC) @ CERN - Open Source developer.

Jan '21 - Jun '21

Visual Computing Group (VCG), Harvard - Visiting Researcher.

Jan '20 - Dec' 20

APPCAIR Lab & TCS Research - Undergratuate Research Assistant.

Jul '20 - Sep '20

ECMWF - Machine Learning Research Intern.

May '20 - Jun '20 (Directi) - Software Engineering Intern.

May '19 - Jul '19

Bank of Maharashtra, Head Office - Summer SWE Intern.

Feb '19 - Apr '19

Pixxel - Machine Learning Intern.

Spring '21

BITS G513: Meta-Learning - Teaching Assistant (TA) [Graduate course]

Fall '20

BITS F464: Machine Learning - Teaching Assistant (TA).

Summer '20

iXperience: Data-Science Program - Teaching Assistant (TA).

Spring '20

BITS F464: Machine Learning - Teaching Assistant (TA).

Fall '19

Technology Incubator Programme, BITS Pilani - Project Mentor.

Summer '19

Quark-QSTP: Introduction to Machine Learning - Instructor.

Normalizing Flows for Fast Detector Simulation
Supervisors : Prof. Sergei V. Gleyzer, Prof. Harrison B. Prosper, & Prof. Michelle Kuchera.
[ Blog ]

Research on using Normalizing Flows to simulate detector response of particle jet deposits. Using the generative capabilities of recent Normalizing Flow models like MAF and BNAF to model conplex jet distributions encountered when simulating detector deposits.

Compositional Reasoning and Visual Understanding on Videos
Supervisors : Prof. Ashwin Srinivasan, Dr Shirish Karande

We explore compositional reasoning on the CATER dataset by learning action embeddings and by using object centric representations.

[ Full Size ]

[ Full Size ]

[ Full Size ]

[ Full Size ]

Check out the Independent Demo Page for more information!

[ Go to Independent Demo Page ]

A Review of Some Techniques for Inclusion of Domain-Knowledge into Deep Neural Networks
Nature Scientific Reports
Tirtharaj Dash, Sharad Chitlangia, Aditya Ahuja, Ashwin Srinivasan
[ Arxiv ]

Abstract: We present a survey of ways in which existing scientific knowledge are included when constructing models with neural networks. The inclusion of domain-knowledge is of special interest not just to constructing scientific assistants, but also, many other areas that involve understanding data using human-machine collaboration. In many such instances, machine-based model construction may benefit significantly from being provided with human-knowledge of the domain encoded in a sufficiently precise form. This paper examines the inclusion of domain-knowledge by means of changes to: the input, the loss-function, and the architecture of deep networks. The categorisation is for ease of exposition: in practice we expect a combination of such changes will be employed. In each category, we describe techniques that have been shown to yield significant changes in the performance of deep neural networks.

Anomaly Detection in Streaming Time Series Data
Supervisor : Dr Peter Dueben
[ Project | Presentation ]

ECMWF (Europe's largest meteorological research institute) acts a data vendor to several clients, providing massive amounts of meteorological data to them. We work on building intelligent Anomaly Detection systems capable of monitoring the log files produced by these services for sudden disruptions and failures.

This work was supported by a grant of £5,000 from ECMWF.

[ Demo | Full Size ]

[ Demo | Full Size ]

Check out the Independent Demo Page for more information!

[ Go to Independent Demo Page ]

Developing a framework to model solutions for Bongard Problems
Supervisors : Prof. Ashwin Srinivasan, Dr Lovekesh Vig

The Bongard problems were introduced by Mikhail Moiseevich Bongard in 1967, in his classic pattern recognition book. We use the DeepProbLog framework to model solutions to these problems by evaluating different hypotheses and their respective likelihoods.

Detecting schizophrenia using Electroencephalography Signals
Supervisor : Prof. Amalin Prince

Schizophrenia is a mental disorder whith symptoms including hallucinations and episodes of psychosis. We develop a Deep Learning Pipeline for automated Schizophrenia detection using abnormalies in brain-wave patterns captured through EEG Data.

Implementing Spiking-Time Dependent Plasticity on SpineCreator
Supervisor : Prof. Basabdatta Sen Bhattacharya
[ Report | Poster ]

Spike-Timing Dependent Plasticity, or STDP is the proposed theory which aims to relate temporal spike differences to changes in synaptic weights between participating neurons. We explore how well STDP works on a model of the Basal ganglia, using the Izhikevich neuron, using SpineCreator to model the underlying networks.

BlackSwan - Realtime Streaming Anomaly Detection on Time Series
[ Project ] [ Presentation ]

Developed a pipeline for Deep Time Series Anomaly Detection on Server Log files. Integrated several State of the Art Anomaly Detection and Forecasting algorithms with a Realtime plotting framework into a python package. This work was funded by ECMWF's open source program - ESoWC.

Emotion Recognition from Audio Signals
[ Code ] [ HTML ]

Developed a Deep Learning pipeline for Emotion recognition and classification using speech data, on the MELD Dataset. Classified emotions across various emotions : [Disgust, Fear, Neutral, ...] across a highly unbalanced data sample. Used Mel-frequency cepstral coefficients (MFCCs) to form speech representations.

Memotion Sentiment Analysis
[ Code ] [ Preprocessing & Model ]

Integrated deep text and image processing models to build a Multimodal Sentiment Analysis system that classified emotions on Internet Memes across different categories. This work was done on the Sem-Eval dataset.

Conway's Game of Life
[ Code ]
Demos: Still Life, Oscillators, Acorn Spread, Engine Spread, Guns, Pulsars, Ships Simple, Ships Collision, Ships Destroyed.

A modular implementations of Conway's Game of Life in Python with common patterns like Still lifes, Oscillators, and Spaceships.

Signature Verification using Siamese Networks
[ Code ] [ HTML ]

Developed a Siamese Neural Net that performed few shot signature verification. Taking a few sample signatures of a person, the model predicted whether the input 'query' signature fas forged or authentic.

Visualizing Genomic data
[ Code ] [ HTML ]

Worked with genomic data from different geographical locations, plotting it after dimensionality reduction. Demonstrated relations between geographic origin and DNA structure by generating different plots and identifying aggregations.

Image Generation using GANs
[ Code ] [ HTML ]

Developed a Generative Adversarial Network (GAN) to generate new instances of the CIFAR Dataset.

Generating Word Embeddings
[ Code ] [ HTML ]

Generated word embeddings using GloVe, and the Large Movie Review Dataset. Visualized the obtained embeddings on a 2D graph using Principal Component Analysis (PCA), and checked the obtained embeddings for semantic coherence.

[ Toggle All ] [ Template ]