Winstead Xingran Zhu
Hi, I’m Winst.
I love Natural Language Processing, Machine Learning, and Statistics.
I also love elephants.
My co-adopted elephant at Nairobi Nursery Unit is called Roho.
You can help more orphan elephants at Sheldrick Wildlife Trust.
Course topics include machine learning and deep learning, neural networks, software engineering and project management, sentiment analysis, language modelling, sequential models and Transformers, machine translation, speech technology, information retrieval, object-oriented programming and design, data structures, algorithm and complexity, etc.
Course topics include statistical learning for risk modelling, statistical inference, probability and statistics, mathematical analysis and methods, stochastic models, linear models and forecasting, financial mathematics, financial derivatives, risk theory, credibility theory and loss distribution, etc.
Physics & applied psychology.
Podcast audio discovery and exploration.
Designing an automatic podcast episode trailer generation system that takes an arbitrary podcast episode as input, analyzes its transcript and audio, and detects hotspots (e.g. strong emotions, high arousal segments, laughter, argument) from the episode, which are then used for generating a 1-minute trailer for the episode.
5LN440 Information Retrieval
5LN715 Advanced Programming for Language Technologists
5LN710 Natural Language Processing
5LN429 Programming for Language Technologists I
Wokring for DECRYPT project. Transcribing historical manuscripts, organizing symbol sets for transcription, editing and re-compiling previous transcriptions, using ML tools for symbol image recognition and manuscript annotation.
Testing and validating time-series bank-failure prediction models, identifying patterns in large-scale market data, performing risk exposure quantification, providing the client bank with proactive solutions to enhance business model robustness, predicting the client bank’s credit rating, market capitalization and solvency position.
Identifying patterns from previous insurance market trends, predicting impacts of future scenarios on interest rates and insurance claim levels using Monte Carlo simulation, constructing time-series ARMA-model forecasting tools for longitudinal trend analysis of post-market life insurance coverage
La terre est bleue comme une orange
Jamais une erreur les mots ne mentent pas
Ils ne vous donnent plus à chanter
Au tour des baisers de s'entendre
Les fous et les amours
Elle sa bouche d'alliance
Tous les secrets tous les sourires
Et quels vêtements d'indulgence
À la croire toute nue.
Les guêpes fleurissent vert
L'aube se passe autour du cou
Un collier de fenêtres
Des ailes couvrent les feuilles
Tu as toutes les joies solaires
Tout le soleil sur la terre
Sur les chemins de ta beauté.
(L'amour la poésie, 1929)