- 293
- 169 272
Alexander Jung
Austria
เข้าร่วมเมื่อ 18 ก.ย. 2009
Welcome to my personal TH-cam channel! I'm Alexander Jung, a tenured Associate Professor for Machine Learning at Aalto University in Finland. Delving into the mathematical underpinnings of machine learning is my passion, both in research and teaching capacities. Here, you'll find an array of videos covering diverse machine learning topics, spanning from fundamental theoretical principles to practical applications of various techniques.
Connect with me on LinkedIn ("aljung" with 30k followers) and GitHub ("alexjungaalto").
Connect with me on LinkedIn ("aljung" with 30k followers) and GitHub ("alexjungaalto").
Determining User Preference Profiles from Email And User Engagement Data
Master thesis presentation by Vilhelm Toivonen.
มุมมอง: 30
วีดีโอ
Machine Learning Kinship Verification
มุมมอง 41หลายเดือนก่อน
Aleksandra Shuianova presents her MSc thesis work.
Object Detection for Security Systems
มุมมอง 42หลายเดือนก่อน
Artem Rodimov presents his Master thesis completed at Aalto University.
Backdoor Attacks on Transformer-Based Weather Forecast
มุมมอง 402 หลายเดือนก่อน
Rudolf Mård presents his master thesis about backdoor attacks on transformer-based regression methods.
Traffic Accident Forecasting with Deep Learning
มุมมอง 804 หลายเดือนก่อน
Traffic accidents are a significant concern for urban safety, and accurate predictions of accidents can help prevent them. The city of Helsinki aims to reduce the number of traffic accidents significantly. In his master's thesis at Aalto University, Akseli Manninen applied deep learning to predict weekly traffic accidents in the city area of Helsinki. @aaltouniversity #deeplearning #roadsafety ...
Trustworthy Object Detection
มุมมอง 745 หลายเดือนก่อน
Raili Virtanen presents her master thesis on trustworthy object detection in industrial environments.
CS-E4740 Data and Model Poisoning
มุมมอง 3425 หลายเดือนก่อน
This lecture explores various data and model poisoning attacks on machine learning systems, with a special focus on federated learning. Dive into the vulnerabilities and understand the key challenges in safeguarding these technologies from such threats. Read more in Section 10 of the lecture notes: github.com/alexjungaalto/FederatedLearning/blob/main/material/FL_LectureNotes.pdf Coding assignme...
CS-E4740 Privacy Protection in FL
มุมมอง 2.3K5 หลายเดือนก่อน
This lecture discusses quantitative measures for - and techniques for minimization of - privacy leakage in federated learning systems. It is based on Section 9 of the lecture notes: github.com/alexjungaalto/FederatedLearning/blob/main/material/FL_LectureNotes.pdf Corresponding coding assignment can be found on my @GitHub github.com/alexjungaalto/FederatedLearning/blob/main/material/Assignments/...
CS E4740 FL at Nokia Part II
มุมมอง 1116 หลายเดือนก่อน
Guest talk about standardization for federated deep learning within our course CS-E4740 Federated Learning. #machinelearning #datascience @Nokia
CS-E4740 Federated Learning at Nokia - Part I
มุมมอง 1266 หลายเดือนก่อน
Guest lecture by Nokia experts within our course CS-E4740. @Nokia #machinelearning #distributedsystems
CS-E4740 Privacy-Preserving Data Processing
มุมมอง 2966 หลายเดือนก่อน
Prof. Luca Vassio from @ discusses some important techniques for protecting the privacy of humans within data processing systems (such as federated learning methods). The slides of the lecture can be found here: github.com/alexjungaalto/FederatedLearning/blob/main/material/slidesvassio.pdf #machinelearning #gdpr #privacymatters @flowerlabs @F-Secure
CS-E4740 Explainable Federated Learning
มุมมอง 1536 หลายเดือนก่อน
We discuss a simple measure for the subjective explainability of a trained model. This measure is then used as regularizer for the personalized model training, resulting FL systems that not only learn tailored personalized models but also ensure tailored explainabilty for each user. Read more about subjective explainability within FL systems in Section 8 of the lecture notes: github.com/alexjun...
CS-E4740 Human Agency and Oversight in Federated Learning
มุมมอง 316 หลายเดือนก่อน
Extract from the lecture "Trustworthy FL" within the course CS-E4740 Federated Learning, taught during spring 2024 at Aalto University. Read more in Section 8 of the lecture notes here: github.com/alexjungaalto/FederatedLearning/blob/main/material/FL_LectureNotes.pdf
Humans must regulate Technology
มุมมอง 186 หลายเดือนก่อน
We compare the regulation of federated learning systems with the (somewhat successful) regulation of how we use motorized vehicles ("cars").
CS E4740 From Linear to Non Linear Models
มุมมอง 776 หลายเดือนก่อน
CS E4740 From Linear to Non Linear Models
CS-E4740 Guest Talk from Javier Fernandez-Marques (Flower Labs)
มุมมอง 3256 หลายเดือนก่อน
CS-E4740 Guest Talk from Javier Fernandez-Marques (Flower Labs)
CS E4740 Gradient Descent for Non-Parametric Models?
มุมมอง 1.6K7 หลายเดือนก่อน
CS E4740 Gradient Descent for Non-Parametric Models?
CS-E4740 Lecture "FL Design Principle"
มุมมอง 5577 หลายเดือนก่อน
CS-E4740 Lecture "FL Design Principle"
Three Perspectives of Regularization in Machine Learning
มุมมอง 727 หลายเดือนก่อน
Three Perspectives of Regularization in Machine Learning
It is enjoyable and easy to listen to your lectures, thank you for uploading them. The pacing is perfect for me.
Glad you like them!
hello sir, do you have the pdf she teaches? i want it where can i find it
thanks
Herrlich
Thanks a lot sir for providing these lectures they will be really helpful for my projects
Top
Plz cm my favorite new video🎉🎉🎉
Wow wow wow wow wow wow wow wow wow wow wow🎉🎉🎉🎉
Thank you for the video!
An amazing lecture sir. I am really inspired by your way of teaching. Thank you very much for this lecture sir.
So nice of you
CORRECTION: I have used a wrong choice for the vector q (around 24:05). The correct choice for vector q is presented in Eq. 2.7 of the lecture notes (github.com/alexjungaalto/FederatedLearning/blob/main/material/FL_LectureNotes.pdf). Note that Eq. 2.7 uses also a normalization by the sample size for both, matrix Q and vector q.
During the end of this lecture I discuss the computational complexity of inverting and multiplying matrices. You can read more about this topic e.g. here en.wikipedia.org/wiki/Computational_complexity_of_mathematical_operations and here en.wikipedia.org/wiki/Computational_complexity_of_matrix_multiplication
Hello, Recently I explore your TH-cam channel and noticed you've been facing some challenges with your channel's growth, particularly in terms of views. I'm here to lend a hand and offer my expertise in video SEO, promotion, and monetization to give your channel the boost it deserves. Together, we can fine-tune your video titles to grab attention, use engaging descriptions,use rank tag, and design thumbnails that invite clicks. Your content isn't optimized and SEO friendly that's why you didn't get actual result of your videos.If you do the above things properly you will see the amazing result of your content.If you're up for a chat, I'd love to hear more about your goals and challenges, and also if you want then I can send your channel audit file free. Looking forward to the opportunity to help your channel thrive. Thanks and Regards Mukta.
it would be very cool to bring the camera closer to the board
Is there a code base linked to this paper/project ?
With this mean the Compressed Federated Distillation that I‘m interested in.
Thank you!! The visualization is awesome!
😳 Promo*SM
How to find optimal number of clusters in fuzzy c means
Very nice
really interesting....great professor 👏👏👏👏
Thank you for your sharing. Could you please suggest some ideas about applying FL in Computer Vision?
www.nature.com/articles/s41598-022-05539-7
Very good
Cool for who did you do this project? Did you cover all roadmarkings?
This was a collaboration with city planners from #helsinki.
Thanks man. I use the online draft of your book for ERM. It's easier to read than Shalev-Schwartz and Ben-David
che corso meraviglioso, scoperto tardi (solo oggi) ma lo seguirò sicuramente!!!! Bravo Bravo Bravo!!!!!👏👏👏👏
Fascinating! 🌟
What’s the red pie?
Chocolate
What are the sugared ball-candy called please?
Fazer Fruit Jellies
Are you skiing?
Yes
Shining
😃👍🏻
Thanks for sharing. Video is quite small at the corner of your desktop.
Thx. Yes, the recording was sub-optimal. I mainly uploaded the recording to have access to the TH-cam annotation tool :-)
Thank you so much
can you provide the code of this scenario?
Could you please provide the link of this dataset?
Can you provide the code for the same?
Hello Sir i have this question i dont what to write. Plz can you help me Discuss, in relation to satellite-based remote sensing, the radiative transfer model of the visible portion of the electromagnetic spectrum in the ocean, including in your answer the various types of scattering and interactions between electromagnetic radiation from the sun and atmospheric constituents.
So glad I found this video!! Do not waste your time = P-R-O-M-O-S-M!
Excellent piece of work Pasi. Another great example how tech can help International Development sector.
Ausgezeichnet.
Thanks! Sadly, Kimi will never win again :(
crystal clear explanation, thank you so much!
Can you please share the code
This is a brilliant introduction. Where can I watch further lectures?
Great lecture Andi, thanks again!
thats sound really intresting. What do you mean with subject? So do you mean each subject is a different thought. And that you can differenciate between the thoughts with the accuracy you showd in the result part?
Very good questions Leon. If not already done, pls contact directly Jaakko.
😍
I'm a professional data scientist and i'm learning tons from these video's, thanks alot!
thank you werner herzog
Haha. Did we meet at University of Edinburgh a while ago ? I remember that some Post-Doc compared my English to that of Werner Herzog.
@@alexjung111 No I'm a student in the united states haha. I did enjoy your voice tho
Hey do you have that code?