Published:
Let’s say you are tracking a drone with a radar. You try to measure the position with the sensor, but the measurement is inaccurate and doesn’t capture its true position. You have another approach, model the motion of the drone, but your model is not perfect, due to environmental factors such as wind. You now have two inperfect approaches, so how do you go about tracking the motion of a drone?
Published:
In the field of Gravitational-Wave astronomy, I see a transition from traditional stochastic sampling methods such as Monte Carlo Markov Chains and nested sampling towards something called normalising flows. As our inference problems become harder due to the increasing volume of data, we need faster inference methods. In this blog post, we will go through what is a normalising flow and can it solve inference problem that are intractably long for traditional methods.
Published:
I have used Gaussian process and it being used extensively in fields such as astrophysics. They are very useful and flexible models that can be applied to timeseries data. However, I often find it difficult to explain a Gaussian process (GP) to a non-expert in a short amount of time. While it is easy to use many of the GP libaries available, wrapping your head around the theoretical concepts can be challenging.
Published:
Lately, I have wondered how people forecast the stock market? How do people in the finance industry make so much money by predicting the future?
Published:
Recently, I came across this kaggle dataset for student stress suvery conducted nationwide. There are 20 factors where participants must give a numerical score to rate how prevalent these factors are. This dataset is pretty interesting as stress is something that affects us all, and understanding what factors are correlated with increased stress levels can improve our quality of life.