Skip to content

Latest commit

 

History

History
35 lines (31 loc) · 4.34 KB

README.org

File metadata and controls

35 lines (31 loc) · 4.34 KB
Raw

Group meeting schedule, Fall 2022

Time 2:30-3:30pm on Friday afternoons in Building 90 (SICCS), room 223, http://whenisgood.net/3c2yh5b/results/4ainmhj

What to present?

  1. your research.
  2. somebody else’s paper.
  3. some useful software.
  • Sep 9: Toby, Title: Time complexity analysis of recently proposed algorithms for optimal changepoint detection. Slides: https://github.com/tdhock/PeakSegFPOP-paper/blob/master/2022-09-28-tucson-slides.pdf Abstract: Several dynamic programming algorithms have been recently proposed for efficiently solving various problems related to optimal changepoint detection in large data sequences measured over space or time. In this talk we will discuss three recent papers which each provide a theoretical or empirical time complexity analysis. Our time complexity analysis shows that these linear and log-linear time algorithms can be used for analysis of huge data sequences, with ten million observations or more, which are now common in areas such as genomics.
  • Sep 16: Daniel, Title: Ghana and me. This is a presentation about myself and the culture of the West Arican country, Ghana. Slides: https://github.com/EngineerDanny/microbe-network-research/raw/main/Fall-2022/week-1/Ghana-and-me.pptx. My second presentation will be about the research paper “Network Analysis Methods for studying microbial communities : A mini review”. Slides: https://github.com/EngineerDanny/microbe-network-research/raw/main/Fall-2022/week-1/network-analysis-presentation.pptx
  • Sep 23: Kyle
  • Sep 30: Jadon, Title: Optimizing Looping ROC Curves. We’ll discuss Area Under ROC Curves, and showcase how real world data can lead to loops within our ROC curves. We will also go over a surrogate loss function for AUC, called Area Under Min(FP, FN) (AUM), that we can use to optimize these looping ROC curves. Finally, we’ll look at ongoing research for efficient AUM optimization: replacing the quadratic time exact line search with a log-linear time approximation. Links: Slides, AUM paper
  • Oct 7: Cam practice exam talk.
  • Oct 14: Group photo? Bilal Aslam, Title: on-road Carbon Emission estimation using a regression model. AbstractL In my model I tried to predict the Annual Average Daily Traffic (AADT) from Population Density, Sentinel-2 Optical Bands, Road Density and Nighttime Light Data. I will be presenting details about my datasets, my model and results. I understand my work might not be presentable or accurate to present in your lab at the moment, but it will help me to understand my weaknesses and can give me a way forward. Please let me know if you have any questions or need more information.
  • Oct 21: Toby out of town.
  • Oct 28: Gabi.
  • Nov 4: Daniel, Title: LASSO Regression modeling on prediction of medical terms among seafarers’ health documents using tidy text mining. I will be presenting on the literature review of the research paper. Links: Slides.
  • Nov 11: Veteran’s Day Holiday.
  • Nov 18: Bilal practice PHD exam 1. I will be presenting my Comprehensive Exam presentation. It will be 25 minutes long. It’s basically the overview of my proposed idea of research, why it’s important, what research gap I am trying to fill and what will be my research objectives.
  • Nov 25: Thanksgiving holiday.
  • Dec 2: Kyle practice master thesis defense- Abstract: This research study will propose functional representations of the square and square-hinge loss. The study will contain the necessary background to understand the problem, theoretical analysis of the proposed algorithm, empirical studies to demonstrate their value, and a discussion of future work. The results of this study will show that it is now possible to use the square/squared hinge loss with gradient descent learning and large batch sizes to achieve better model performance metrics. Slide Deck