Friday, April 29, 2022

4/29/22: More reading

         Agenda & Minutes

  • Welcome to the 14th meeting.
  • Any updates/news/inputs/comments?
  • Readings, viewings, etc.: 
    • Sources to read/view in more depth.
      • https://e2eml.school/transformers.html: "Transformers From Scratch." We read up to "First order sequence model" so next time we will start there.
        • Remember that the vote was 4 3/8 out of 5 for this document, which can always be updated as we progress through it.
The meeting ended here.
  • Potential future readings that we have assessed in earlier meetings.
    • In a previous meeting we read through the fourth paragraph of  https://www.marktechpost.com/2022/03/07/an-introduction-to-saliency-maps-in-deep-learning/. Previously we read the fifth paragraph, then voted on the priority for reading more of it. Vote: 3.67 out of 5.
    • CNN basics: https://towardsdatascience.com/the-most-intuitive-and-easiest-guide-for-convolutional-neural-network-3607be47480. We previously read 2 paragraphs. Read more? Vote was 3.6 out of 5.
    • https://www.youtube.com/watch?v=BolevVGJk18. This introduces Jonschkowski, Brock, Learning State Representations with Robotic Priors. Should we try the first paragraph(s) of the paper? Vote was 3.6 out of 5.
    • Ni et al., Learning Good State and Action Representations via Tensor Decomposition, https://arxiv.org/abs/2105.01136. We read the title and 1st sentence. Vote to read more was 4 out of 5.
    • Brooks, R., 2017, Seven Deadly Sins of AI Prediction, in serveinfo\AIstudyGroup. Vote was 2.6 out of 5.
  • We can read/view the first paragraph/minute or so of different sources, assessing each whether to go over it in more depth. To assess each one, vote: Should we read/view more of this? 5=strongly agree, 4=agree, 3=neutral, 2=disagree, 1=strongly disagree.
        • https://en.wikipedia.org/wiki/Markov_decision_process. 
        • MM suggests explainable AI as a reading/discussion topic. 

      at No comments:

      Friday, April 22, 2022

      4/22/22: News and a reading

              Agenda & Minutes

      • Welcome to the 13th meeting.
      • Any updates/news/inputs/comments?
      • Readings, viewings, etc.: 
        • Sources to read/view in more depth.
          • https://e2eml.school/transformers.html: "Transformers From Scratch." We previously read through the 2nd paragraph. This time we read up to the phrase "The previous two examples show how dot products" so we will continue there next time. The vote was 4 3/8 out of 5 which can always be updated as we progress through it and see how it goes.
      The meeting ended here.
        • Potential future readings that we have assessed.
          • In a previous meeting we read through the fourth paragraph of  https://www.marktechpost.com/2022/03/07/an-introduction-to-saliency-maps-in-deep-learning/. Previously we read the fifth paragraph, then voted on the priority for reading more of it. Vote: 3.67 out of 5.
          • CNN basics: https://towardsdatascience.com/the-most-intuitive-and-easiest-guide-for-convolutional-neural-network-3607be47480. We previously read 2 paragraphs. Read more? Vote was 3.6 out of 5.
          • https://www.youtube.com/watch?v=BolevVGJk18. This introduces Jonschkowski, Brock, Learning State Representations with Robotic Priors. Should we try the first paragraph(s) of the paper? Vote was 3.6 out of 5.
          • Ni et al., Learning Good State and Action Representations via Tensor Decomposition, https://arxiv.org/abs/2105.01136. We read the title and 1st sentence. Vote to read more was 4 out of 5.
          • Brooks, R., 2017, Seven Deadly Sins of AI Prediction, in serveinfo\AIstudyGroup. Vote was 2.6 out of 5.
        • We can read/view the first paragraph/minute or so of different sources, assessing each whether to go over it in more depth. To assess each one, vote: Should we read/view more of this? 5=strongly agree, 4=agree, 3=neutral, 2=disagree, 1=strongly disagree.
            • https://en.wikipedia.org/wiki/Markov_decision_process. 
            • MM suggests explainable AI as a reading/discussion topic. 

          at No comments:

          Friday, April 15, 2022

          4/15/22: Presentations - AM and SN

                 Agenda & Minutes

          • Welcome to the 12th meeting.
          • Presentations.
            • 1. AM
            • 2. SN
            • We discussed the implications of these presentations.
          The meeting ended here.
          • Any updates/news/inputs/comments?
          • Readings, viewings, etc.: 
            • Sources to read/view in more depth.
              • https://e2eml.school/transformers.html: "Transformers From Scratch." We previously read through the 2nd paragraph. This time we will read the 3rd paragraph and vote on continuing with the document in the future. Vote was 4 3/8 out of 5.
            • Potential future readings that we have assessed.
              • In a previous meeting we read through the fourth paragraph of  https://www.marktechpost.com/2022/03/07/an-introduction-to-saliency-maps-in-deep-learning/. Previously we read the fifth paragraph, then voted on the priority for reading more of it. Vote: 3.67 out of 5.
              • CNN basics: https://towardsdatascience.com/the-most-intuitive-and-easiest-guide-for-convolutional-neural-network-3607be47480. We previously read 2 paragraphs. Read more? Vote was 3.6 out of 5.
              • https://www.youtube.com/watch?v=BolevVGJk18. This introduces Jonschkowski, Brock, Learning State Representations with Robotic Priors. Should we try the first paragraph(s) of the paper? Vote was 3.6 out of 5.
              • Ni et al., Learning Good State and Action Representations via Tensor Decomposition, https://arxiv.org/abs/2105.01136. We read the title and 1st sentence. Vote to read more was 4 out of 5.
              • Brooks, R., 2017, Seven Deadly Sins of AI Prediction, in serveinfo\AIstudyGroup. Vote was 2.6 out of 5.
            • We can read/view the first paragraph/minute or so of different sources, assessing each whether to go over it in more depth. To assess each one, vote: Should we read/view more of this? 5=strongly agree, 4=agree, 3=neutral, 2=disagree, 1=strongly disagree.
                  • https://en.wikipedia.org/wiki/Markov_decision_process. 
                  • MM suggests explainable AI as a reading/discussion topic. 

              at No comments:

              Friday, April 8, 2022

              4/8/22: Read more beginnings and vote on continuing with them

                    Agenda & Minutes

              • Welcome to the 11th meeting.
              • Any updates/news/inputs/comments?
              • Readings, viewings, etc.: 
                • We can read/view the first paragraph/minute or so of several sources, and then vote to pick one to do in more depth. Here is a way to vote to pick the next reading:
                  • Vote whether to read more, using a scale 1-5: 
                    • Should we read/view more of this? 
                      • 5=strongly agree, 4=agree, 3=neutral, 2=disagree, 1=strongly disagree.
                  • Repeat the process for another article.
                  • After going through a few articles this way, we can pick the one with the best voting result to read more from.
                1. Last time we read through the fourth paragraph of  https://www.marktechpost.com/2022/03/07/an-introduction-to-saliency-maps-in-deep-learning/. Previously we read the fifth paragraph, then voted on the priority for reading more of it. Vote: 3.67 out of 5.
                2. CNN basics: https://towardsdatascience.com/the-most-intuitive-and-easiest-guide-for-convolutional-neural-network-3607be47480. We previously read 2 paragraphs. Read more? Vote was 3.6 out of 5.
                3. https://e2eml.school/transformers.html: "Transformers From Scratch." We previously read through the 2nd paragraph. This time we will read the 3rd paragraph and vote on continuing with the document in the future. Vote was 4 3/8 out of 5.
                4. https://www.youtube.com/watch?v=BolevVGJk18. This introduces Jonschkowski, Brock, Learning State Representations with Robotic Priors. Should we try the first paragraph(s) of the paper? Vote was 3.6 out of 5.
                5. Ni et al., Learning Good State and Action Representations via Tensor Decomposition, https://arxiv.org/abs/2105.01136. We read the title and 1st sentence. Vote to read more was 4 out of 5.
                6. Brooks, R., 2017, Seven Deadly Sins of AI Prediction, in serveinfo\AIstudyGroup. Vote was 2.6 out of 5.
              We got to here.
              • Featured resource: Short and long videos - 
              • Some quantum computing references we could read as needed (from VW):
                • - Quantum crossing threshold (free): https://www.nature.com/articles/s41586-021-04273-w
                • - Crossing threshold in silicon: https://www.nature.com/articles/s41586-021-04182-y
                • - Three-qubit donor processor in Si: https://www.nature.com/articles/s41586-021-04292-7

              at No comments:

              Friday, March 25, 2022

              4/1/22: Read paragraphs and decide whether to read more

                   Agenda & Minutes

              • Welcome to the 10th meeting.
              • Any updates/news/inputs/comments?
              • Readings, viewings, etc.: 
                • We can read/view the first paragraph/minute or so of several sources, and then vote to pick one to do in more depth. Here is a way to vote to pick the next reading:
                  • Vote whether to read more, using a scale 1-5: 
                    • Should we read/view more of this? 
                      • 5=strongly agree, 4=agree, 3=neutral, 2=disagree, 1=strongly disagree.
                  • Repeat the process for another article.
                  • After going through a few articles this way, we can pick the one with the best voting result to read more from.
                • Last time we read through the fourth paragraph of  https://www.marktechpost.com/2022/03/07/an-introduction-to-saliency-maps-in-deep-learning/. This time we read the fifth paragraph, then voted on the priority for reading more of it. Vote: 3.67 out of 5.
                • CNN basics: https://towardsdatascience.com/the-most-intuitive-and-easiest-guide-for-convolutional-neural-network-3607be47480. We read 2 paraagraphs. Read more? Vote was 3.6 out of 5.
                • https://e2eml.school/transformers.html: "Transformers From Scratch." We read through the 2nd paragraph. Next time we will read the 3rd paragraph and vote on continuing with the document in the future. 
              We ended here.
                  • We have read through section 3 so we could start with 3.1 next time we look at it.
                • Featured resource: Short and long videos - 
                • Some quantum computing references we could read as needed (from VW):
                  • - Quantum crossing threshold (free): https://www.nature.com/articles/s41586-021-04273-w
                  • - Crossing threshold in silicon: https://www.nature.com/articles/s41586-021-04182-y
                  • - Three-qubit donor processor in Si: https://www.nature.com/articles/s41586-021-04292-7

              at No comments:

              Friday, March 11, 2022

              3/18/22: Browsing for concepts

                  Agenda & minutes:

              • Welcome to the 9th meeting.
              • Any other updates/news/inputs/comments?
              • Should we try assigning a reading one week, and then discussing it the next week, or stay with the current format where we read together and discuss, a little bit at a time?
              • Future meetings:
                • Next meeting(s): Any suggestions or requests? 
              • Readings, viewings, etc.: 
                • We read the first three paragraphs of https://www.marktechpost.com/2022/03/07/an-introduction-to-saliency-maps-in-deep-learning/ last and could potentially continue with it next time.
                • Should we try reading the first paragraph or so of several sources, and then pick one to do in more depth? Or go with picking something and trying to get through it? We could vote on this.
              • Future reading possibilities:
                • CNN basics: https://towardsdatascience.com/the-most-intuitive-and-easiest-guide-for-convolutional-neural-network-3607be47480
                • https://e2eml.school/transformers.html: "Transformers From Scratch"
                • https://www.youtube.com/watch?v=BolevVGJk18
                • Jonschkowski, Brock, Learning State Representations with Robotic Priors, on disk
                • Ni et al., Learning Good State and Action Representations via Tensor Decomposition, on disk
                • MM would like to step us through some of the resources available from NVIDIA.
                • MM suggests explainable AI as a reading/discussion topic.
                • MM suggests https://www.youtube.com/watch?v=4Bdc55j80l8&ab_channel=The A.I.Hacker-MichaelPhi as a transformer video.
                • 2021 Turing Award lecture paper: https://dl.acm.org/doi/pdf/10.1145/3448250
                • Anticipative Video Transformer, https://facebookresearch.github.io/AVT/?fbclid=IwAR1RurSM33v8baN10H9JCX_dvVNtscydsLupaB8NMgKOmNIPjIwD3XO2vOA.
                • "Deep learning—a first meta-survey of selected reviews across scientific disciplines, their commonalities, challenges and research impact," https://peerj.com/articles/cs-773.
                • We read the abstract. It is not clear whether we should continue reading material from it. Any opinions/thoughts/comments?
                • "Attention is all you need," https://proceedings.neurips.cc/paper/2017/file/3f5ee243547dee91fbd053c1c4a845aa-Paper.pdf. Attention is all you need, A Vaswani, N Shazeer, N Parmar… - Advances in neural …, 2017 - proceedings.neurips.cc … Cited by 35,980 
                  • We have read through section 3 so we could start with 3.1 next time we look at it. However this paper seems dense for our current state of understanding, so we'll put it on hold.
                • Featured resource: Short and long videos - 
                • Some quantum computing references we could read as needed (from VW):
                  • - Quantum crossing threshold (free): https://www.nature.com/articles/s41586-021-04273-w
                  • - Crossing threshold in silicon: https://www.nature.com/articles/s41586-021-04182-y
                  • - Three-qubit donor processor in Si: https://www.nature.com/articles/s41586-021-04292-7

              at No comments:

              Friday, March 4, 2022

              3/11/22: Focus on a video about the transformers paper

                 Agenda & minutes:

              • Welcome to the 8th meeting.
              • Any other updates/news/inputs/comments?
              • Future meetings:
                • Next meeting(s): Any suggestions or requests? 
              • Readings, viewings, etc.: 
                • We viewed and discussed https://www.youtube.com/watch?v=4Bdc55j80l8
                • We read the first two paragraphs of https://www.marktechpost.com/2022/03/07/an-introduction-to-saliency-maps-in-deep-learning/ and could potentially continue with it next time.
              Meeting ended here.
              • Future reading possibilities:
                • https://e2eml.school/transformers.html: "Transformers From Scratch"
                • MM would like to step us through some of the resources available from NVIDIA.
                • MM suggests explainable AI as a reading/discussion topic.
                • MM suggests https://www.youtube.com/watch?v=4Bdc55j80l8&ab_channel=The A.I.Hacker-MichaelPhi as a transformer video.
                • 2021 Turing Award lecture paper: https://dl.acm.org/doi/pdf/10.1145/3448250
                • Anticipative Video Transformer, https://facebookresearch.github.io/AVT/?fbclid=IwAR1RurSM33v8baN10H9JCX_dvVNtscydsLupaB8NMgKOmNIPjIwD3XO2vOA.
                • "Deep learning—a first meta-survey of selected reviews across scientific disciplines, their commonalities, challenges and research impact," https://peerj.com/articles/cs-773.
                • We read the abstract. It is not clear whether we should continue reading material from it. Any opinions/thoughts/comments?
                • "Attention is all you need," https://proceedings.neurips.cc/paper/2017/file/3f5ee243547dee91fbd053c1c4a845aa-Paper.pdf. Attention is all you need, A Vaswani, N Shazeer, N Parmar… - Advances in neural …, 2017 - proceedings.neurips.cc … Cited by 35,980 
                  • We have read through section 3 so we could start with 3.1 next time we look at it. However this paper seems dense for our current state of understanding, so we'll put it on hold.
                • Featured resource: Short and long videos - 
                • Some quantum computing references we could read as needed (from VW):
                  • - Quantum crossing threshold (free): https://www.nature.com/articles/s41586-021-04273-w
                  • - Crossing threshold in silicon: https://www.nature.com/articles/s41586-021-04182-y
                  • - Three-qubit donor processor in Si: https://www.nature.com/articles/s41586-021-04292-7

              at No comments:
              Subscribe to: Posts (Atom)