Machine Learning Study Group
|
- Guest presentation: RM on ML for antenna design.
- Transcript of meeting is appended to the end of this page.
- A demo of the real time use of AI to create the Doppler effect interactive animation and other demos will be scheduled as soon as convenient for RM and VW.
- Announcements, updates, questions, etc.?
- The campus has assigned a group to participate in the AAC&U AI Institute's activity "AI Pedagogy in the Curriculum." Our ML Study Group member IU is on it (any news?).
- Anything else?
- Here are the latest on readings and viewings
- Next
we will work through chapter 5:
https://www.youtube.com/watch?v=wjZofJX0v4M. We got up 15:50 and can
start there next time we work on this video! (When sharing the screen,
we need to click the option to optimize for sharing a video.)
- We can work through chapter 6: https://www.youtube.com/watch?v=eMlx5fFNoYc
- We can work through chapter 7: https://www.youtube.com/watch?v=9-Jl0dxWQs8
- Computer scientists win Nobel prize in physics! Https://www.nobelprize.org/uploads/2024/10/popular-physicsprize2024-2.pdf got a evaluation of 5.0 for a detailed reading.
- We can evaluate https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10718663 for reading & discussion.
- Chapter 6 recommends material by Andrej Karpathy, https://www.youtube.com/@AndrejKarpathy/videos for learning more.
- Chapter 6 recommends material by Chris Olah, https://www.youtube.com/results?search_query=chris+olah
- Chapter 6 recommended https://www.youtube.com/c/VCubingX for relevant material, in particular https://www.youtube.com/watch?v=1il-s4mgNdI
- Chapter 6 recommended Art of the Problem, in particular https://www.youtube.com/watch?v=OFS90-FX6pg
- LLMs and the singularity: https://philpapers.org/go.pl?id=ISHLLM&u=https%3A%2F%2Fphilpapers.org%2Farchive%2FISHLLM.pdf (summarized at: https://poe.com/s/WuYyhuciNwlFuSR0SVEt). 6/7/24: vote was 4 3/7. We read the abstract. We could start it any time. We could even spend some time on this and some time on something else in the same meeting.
TRANSCRIPT OF MEETING: Meeting summary (11/01/2024)
R, a Ph.D. student
in electrical and computer engineering, presented on the topic of autonomous
vehicle communication, emphasizing the importance of AI and ML techniques in
her research. She discussed various concepts such as the Doppler effect,
frequency hopping technique, vehicle-to-everything (V2X) communication, and
beamforming, and their applications in wireless communication and
transportation. The meeting also highlighted the potential of machine
learning in beamforming, the significance of energy savings in antenna
technology, and the role of V in demonstrating concepts for R's work.
Next steps
• R to review conference paper and address reviewer comments.
• V and R to schedule a demo of the AI-assisted development process for beam forming visualizations in the coming weeks.
• V to prepare a demonstration of using Claude AI to generate HTML/CSS/JavaScript visualizations for RF concepts.
• R to continue research on applying convolutional neural networks to beam forming for autonomous vehicle communication.
• V to assist R in preparing AI-generated demonstrations for future presentations on beam forming techniques.
Summary
Autonomous Vehicle Communication Techniques
D B handed over the presentation to R. R, a Ph.D. student in electrical and computer engineering, presented on the topic of autonomous vehicle communication, emphasizing the importance of AI and ML techniques in her research. She explained the Doppler effect and its impact on modulation techniques, demonstrating its effects using a cloud artifact. R also highlighted the efficiency of AI in constructing demonstrations, which would previously have taken hours to create manually. She concluded by discussing the sensitivity of different modulation techniques to the Doppler effect.
Frequency Hopping and Beamforming Techniques
R discussed the frequency hopping technique, inspired by Hollywood actress Hedy Lamar, which provides secure communication by making it difficult for enemies to intercept radio signals. She explained the technique using the 2.4 GHz bandwidth of 5G as an example, dividing it into sub-frequency channels and time chunks. R also introduced the concept of beamforming, comparing it to the double slit experiment in physics. She explained the three types of beamforming techniques: analog, digital, and hybrid, and their applications in wireless communication, radar and sonar systems, audio and speech processing, and hearing aids.
Exploring V2X Communication and 5G Technology
R discussed the concept of vehicle-to-everything (V2X) communication, a sub-technology of the Internet of Things. She explained how V2X enables vehicles to connect with their surroundings, making transportation safer and more efficient. She highlighted the potential of V2X to anticipate and alert drivers of potential crashes, and to relay information to other cars, transport infrastructure, pedestrians, and cyclists. R also emphasized the importance of 5G technology for adVced safety and automated driving, as it provides superfast and ultra-reliable connections. She demonstrated the concept using an animation and a video.
Antenna Gain Pattern Steering With AI
R discussed the importance of steering a gain pattern of an antenna towards a desired direction using cloud artifacts and large language models. She highlighted the time-saving aspect of this method, which takes less than a minute compared to traditional methods that take around 4 to 5 hours. R also explained how the gain pattern changes with the addition of more antenna elements and the need for AI to handle the large number of phase and power combinations required. She demonstrated how the 3D pattern of an antenna steers continuously by changing the phase and power, and how this is crucial for autonomous vehicle communication. R concluded by emphasizing the need for AI and ML techniques to address the numerous combinations required for beam steering.
Machine Learning in Beamforming Explained
R discussed the differences between traditional programming and machine learning programming, emphasizing that the former can fit better into a machine learning problem. She explained how a machine learning model can be trained with data and procedures to find solutions to problems, and how it can inferencing to the desired problem. R also discussed the application of machine learning in beamforming, where a model can be trained with a lookup table or codebook of phase and power combinations to steer a beam towards a desired direction. She highlighted the potential of convolutional neural networks in beamforming, which can adapt to increasing scalability of the antenna array. R also compared traditional and machine learning approaches in terms of accuracy, computational efficiency, scalability, flexibility, and practical implementation.
Phased Array Codebook Demo and ML
R presented a codebook demo for a 2-element phased array case, explaining how the antenna gain pattern changes with phase and power variations. She highlighted the potential of using AI techniques to increase the number of elements in the array, which would be time-consuming using conventional methods. R also mentioned her next steps, which include implementing the same codebook construction with machine learning techniques and comparing the results. Yogesh suggested that the traditional and machine learning methods could be complementary, and R agreed, stating that the traditional method's outputs could be used as input data for the machine learning process.
Exploring Beamforming for Energy Efficiency
R and D B discussed the concept of beamforming in antenna technology. R explained that using a single antenna element would result in a circular pattern of radiation, but this could be inefficient as it would expose unnecessary radiation. She suggested that by using multiple antenna elements, they could target their beam towards specific areas, conserving energy and reducing unnecessary exposure to electromagnetic radiation. R estimated that this could result in a 30 to 50% saving of energy, depending on the case. D questioned the significance of saving a fraction of a watt, to which R responded that in the context of 5G base stations, this could translate to substantial energy savings. They also touched on the topic of radiation protection in vehicles, with R mentioning the concept of a Faraday cage.
Doppler Effect Demonstration and Tools
D B requested a demonstration of how the Doppler effect was developed using Claude.ai, which R confirmed was done by running through prompts. V agreed to arrange a separate talk for this demonstration. G and Mariofanna expressed interest in understanding the tools used for the impressive animations, to which R explained that she used Ansys HFSS for electromagnetic simulations and HTML for the graphs. G clarified that R used custom scripts in HTML, CSS, and JavaScript for the graphs.
V's Role in Demonstrations Discussed
R and V discussed the significant role V played in demonstrating concepts for R's work. They agreed on the need for both R's expertise and V's translation skills for future demonstrations. D B proposed scheduling a demo of the process, but R had to prioritize a conference paper deadline. They agreed to schedule a demo for the following week, with R promising to confirm her availability. V emphasized R's substantial behind-the-scenes effort in preparing the presentations.
