宅男视频

Although I see myself primarily as a cosmologist, the chance to learn and advance cutting-edge AI and machine learning techniques made this course very attractive.

Project title:听Anomaly Detection in DESI Spectra using Machine-Learning Dimensionality-Reduction Techniques

Research Group:听Astrophysics

厂耻辫别谤惫颈蝉辞谤(蝉):听听Prof Ofer Lahav

滨苍迟谤辞诲耻肠迟颈辞苍:听

I've come to PhD research relatively late. I听earned my first degree in Theoretical Physics and Computer Science at Cambridge and then worked in finance for a number of years. I first came to 宅男视频 in 2016 to do an MSc in Astrophysics (winning the Harrie Massey Prize). After a bit more finance and family time, I was finally ready to embark on a PhD. I was so excited to be accepted into the CDT DIS programme. Although I see myself primarily as a cosmologist, the chance to learn and advance cutting-edge AI and machine learning techniques made this course very attractive. I also loved the informal yet impressive environment that UCL has to offer. I am hoping that my research will help听push the boundaries in both AI methods and also play a small part in answering some big cosmological questions. As well as science and finance, I鈥檓 also a writer and run a small independent publishing imprint (). At some point, I'm hoping to bring all these interests together!

---

Project description:听听

The Dark Energy Spectroscopic Instrument (DESI) is set to be poised to transform the field of astronomy with its vast data collection capabilities. My research is focused on using machine-learning techniques to sift through this data and identify rare, predicted, and previously unknown objects. In addition, I am exploring the potential of applying anomaly detection concepts and methods to fields beyond astronomy, such as medical imaging and financial fraud detection. This interdisciplinary approach has the potential to significantly advance our understanding and application of anomaly detection.

First year group project:听

鈥淐oreference Resolution鈥� The Guardian 鈥� We worked with the data science team at The Guardian to help them correctly attribute quotes from their archive. Using state-of-the-art natural language processing (NLP) tools we f, we fine-tuned and tested potential models also听annotated a large dataset which they could use going forward. You can read more in this article I co-wrote for .

Placement:听

---

听

听