Deprecated: Creation of dynamic property OMAPI_Elementor_Widget::$base is deprecated in /home2/salonida/public_html/wp-content/plugins/optinmonster/OMAPI/Elementor/Widget.php on line 41

ABOUT ME

SALONI DASH

I am currently a PhD student at the Information School at University of Washington (UW). Previously, I was a Research Fellow at Microsoft Research (MSR) India.  My research interests primarily lie in Misinformation & AI. To know more about my research you can head over to the Research section. 

In my free time I love cooking, reading, spending time with friends, hiking, acrylic painting, playing my ukelele, sketching, reruns of comfort TV shows, Taylor Swift (evermore is better than folklore, sorry) and anything Harry Potter related!. Above all, I love talking to new people, so please reach out to me if  any of this sounds remotely fun! I also sporadically blog about my personal and professional musings (Blog)

RESEARCH

I have been extremely fortunate to have had great mentors who have nurtured and cultivated my research potential. I am currently pursuing my PhD at University of Washington, where I am working with Prof. Emma Spiro on modeling the threats of generative AI for disinformation campaigns.

At Microsoft Research IndiaI was working with Prof. Joyojeet Pal on studying the spread of political polarization, dangerous speech and disinformation during political crises in India. I was also working with Dr. Amit Sharma and Dr. Vineeth Balasubramanian on counterfactual generation for analysing the fairness of machine learning classifiers

I have also had the honor of working with Prof. Isabelle Guyon (co-inventor of Support Vector Machines) and Prof. Kristin Bennett for over two years on generating privacy-preserving  synthetic health data and time-series medical data using  Generative Adversarial Networks (GANs). Please refer to my CV for more details.

My research interests primarily lie in Misinformation & AI. I am interested in understanding and modeling the threats of geenrative models for disinformation operations. I am also interested in leveraging geenrative models for detecting and explaining misinformation and manipulation tactics.  

My current research interests are a culmination of my previous research endeavours. In the Healthcare domain I was interested in addressing the problem of making medical data accesible to the public, without compromising privacy. At MSR, I was also interested in studying & characterizing the role of online influencers in spreading polarization, hate speech, & misinformation on social media platforms, particularly during political crises. I also worked on investigating interpretability and algorithmic fairness through the lens of causal inference.

In general, I am interested in Machine Learning, Natural Language Processing, Social Computing, Human Computer Interaction, and Fairness & Explainability.

Please take a look at my google scholar profile for a comprehensive list.

You can download my regularly updated CV

PROJECTS

INTERSECTION NAVIGATION FOR AUTONOMOUS AGENTS USING DEEP REINFORCEMENT LEARNING

Self-Driving cars are all the rage now. However, one of the tasks that they still struggle with is navigating road intersections. Explore how to teach an autonomous agent to navigate simple intersections using Deep Reinforcement Learning.

MULTI-AGENT REINFORCEMENT LEARNING FOR COOPERATIVE HUNTING SCENARIOS

Wolves are widely known for their cooperative pack hunting. But why do they hunt in packs? Understand this evolution of cooperation by simulating their behavior using Multi Agent Reinforcement Learning .

EXPLORATORY DATA ANALYSIS FOR SURVIVAL PREDICTION ON TITANIC

To Die ot Not To Die? Learn basic exploratory data analysis tools by extracting features for survival prediction on Titanic. Forsee how Jack didn't stand a chance as compared to Rose to make it till the end of the movie.

SIMULATING LOGIC GATES USING MCCULLOCH PITT'S NEURONS

Mcculloch Pitts neuron was the first computational model of a neuron that inspired the neural nets we see today. Discover how this neuron works through an interactive GUI by inputting neuronal weights that will be used to construct logic gates.

CONTACT