Neural Networks and Deep Learning COMP9444 25T1

Course marks released

Posted by Sonit Singh 2 months ago.

Dear Students,

Course marks (hw1, checkpoints, project, and the final exam) are now released and you can view them via the give system or you can also check from below link.
https://cgi.cse.unsw.edu.au/~give/Student/sturec.php

We have scaled project marks based on the peer assessment form. If you find that the total sum of presentation, notebook, and report is not equal to the total proj scores, it means there these are scaled marks. There are many team members who showed concern about other members lower contribution and lack of participation in meetings, writing code, report writing, making slides, and overall working as a team member. We have gone through each of the submitted peer review form and scaled marks.

Please note that mentors and teaching staff have done all checks to make sure we have accurate marking. However, if you think there is a calculation issue or you got zero mark for particular component due to missing file or no submission, please send an email to cs9444@cse.unsw.edu.au within 4 days of this announcement, i.e. 11 May 2025 5pm. Any request after 11 May will not be entertained as we need to finalise marks and upload to central systems.

Kind regards,
Sonit

myExperience Survey closing tomorrow

Posted by Sonit Singh 2 months ago.

Dear Students,

Thanks to everyone those who have filled the myExperience survey for COMP9444. As of today, we got 45% participation rate.

Those who have not submitted yet, could you please fill the myExperience Survey as it is closing today. We need at least 50% participation rate to mean meaningful and statistical sense of the survey.

Please log into https://myexperience.unsw.edu.au/ or to Moodle course page and fill in the survey.

Kind regards,
Sonit

Week 10 Guest Lecture (Tuesday 9-11 am)

Posted by Sonit Singh 2 months ago.

Dear Students,

It is my great pleasure to announce that we have a guest lecture on "Multimodal AI Reasoning and Agentic AI" by Jingying Gao. I think it's really a great opportunity to get some industry insights on Generative AI and Multimodal application from our guest speaker. I would like to see you all in Week 10 lecture. Please find details below:

Title: Multimodal AI Reasoning and Agentic AI

Abstract : In this guest lecture, we delve into the world of multimodal AI and Agentic AI applications. We will explore the mechanics of text-to-image generation models, including the functioning of diffusion models and their fine-tuning using customised datasets. We will discuss the construction of AI applications based on diffusion models and the stable diffusion model's specific tuning methods. The lecture covers the applications of large language models (LLMs) in text generation and will guide us through building a customised chatbot web application using vector databases (VectorDB) and the GPT-4 model. We will also cover how to develop Neuro-Symbolic systems and AI agents to automate and improve workflows. Finally, we will discuss future directions in AI, such as explainable AI, logical reasoning in multimodal tasks, and AI ethics, to inspire students' innovative thinking and ethical considerations.

Bio : Jingying is a research scientist specialising in multimodal and generative AI, with a focus on VQA, multimodal reasoning and explainable AI. She studied her PhD at the AI & Robotics Lab at UNSW. Her research contributions include papers published at NeurIPS, IJCNN, and other conferences. Additionally, she has served as a reviewer for NeurIPS, ECAI, and IJCAI, and co-organized a workshop at IJCNN. She is also a passionate research scientist working in the industry with over six years of active commercial and academic research experience, preceded by years in managerial and technical roles. She has led the development of various AI and Generative AI applications intended for robots aimed at service duties and in financial services.

Kind regards,
Sonit