“Skops” is the name of the framework being actively developed as the link between the scikit-learn and the Hugging Face ecosystems. With Skops, we hope to facilitate essential workflows:
The ability to push scikit-learn models on the Hugging Face Hub
The possibility to try out models directly in the browser
The automatic creation of model cards, to improve model documentation and understanding
The ability to collaborate with others on machine learning projects
Started to poke at the outlines of a religion chapter
4:00 Meeting with Brenda
SBIRs
Working on 3D viz
1:00 LLM Meeting with Isaac and Aaron
2:00 MDA meeting
GPT Agents
Stub out ModelExplorer. Off to a good start. The app framework is up, loading the model and running prompts. Working out how to split the output for analysis
I track my temperature every morning, along with weight and blood pressure. What can I say? I like to record data. Anyway, I came down with Covid after I came back from Spain in July, got Paxlovid and (motly) recovered in about two weeks. I still had a lingering cough, which is now completely gon. Looking back over the data, my temperature was slightly raised for two months. Not exactly Long Covid, but this thing took a looooong time to shake.
SBIRs
9:00 Meeting with Rukan to go over EntityData
Work with Aaron on paper?
Copy over the AI & Weapons paper to Overleaf for updating – done. It turns out that I don’t “own” the paper on Arxiv – Aaron Massey does. Downloaded final from Arxiv and uploaded to Overleaf
GPT Agents
Added twitter query options to TweetDownloader and a way to disable writes to the db for debugging. I’m afraid to test it given that it’s Friday and all
Fix excluded clusters – I think it’s backwards
Not sure if I want to build an App to interact with the model. I think so? If only for quick data gathering
Find and load a model
Calculate percentages
Set prompts, generate text, get embeddings and cluster
Had a good chat with Rukan about kernel methods. Looking at ways to produce information to split out features that the system struggles to learn so that they can be handled differently. This led to thinking about adding that information in as a feature and then testing the reconstruction of the feature (e.g. std dev) vs the calculation based on the output as the loss function.
Talked to Aaron to work out the experiment design
Had to do my MORS expenses AGAIN
9:15 standup
11:30 touch point
Get 3D viz working, and start on texture import and using lat/long. It looks like I can use the utm package. There are pieces missing from the directory that I don’t want to chase. Made a simpler viztest directory
GPT Agents
The computational sociology group had a meeting again! Went over stuff, and I managed to break the Embedding explorter by removing tweet_row from the view. It should be this:
That’s working now
Getting more embeddings and clusters
Struggled a bit with getting the files to save correctly
I think that this could be a really interesting way of debugging vision systems around edge cases. Mask out everything but the attribution area and run separate image-to-text models to see what they identify and then check for consensus.
This also makes me think of what it might mean for human-monitored AI systems. People will need to be trained to quickly identify if the model is behaving poorly and flag it, as opposed to if the decision is correct. It’s more like driving a race car where you have to monitor the performance of the vehicle and adapt to that. You can’t stop racing if the tires are wearing out or you’re running out of fuel. You have to adjust your behavior to optimize the behavior of the system as it is. Which implies that we need simulators of AI systems that break a lot.
Politics has in recent decades entered an era of intense polarization. Explanations have implicated digital media, with the so-called echo chamber remaining a dominant causal hypothesis despite growing challenge by empirical evidence. This paper suggests that this mounting evidence provides not only reason to reject the echo chamber hypothesis but also the foundation for an alternative causal mechanism. To propose such a mechanism, the paper draws on the literatures on affective polarization, digital media, and opinion dynamics. From the affective polarization literature, we follow the move from seeing polarization as diverging issue positions to rooted in sorting: an alignment of differences which is effectively dividing the electorate into two increasingly homogeneous megaparties. To explain the rise in sorting, the paper draws on opinion dynamics and digital media research to present a model which essentially turns the echo chamber on its head: it is not isolation from opposing views that drives polarization but precisely the fact that digital media bring us to interact outside our local bubble. When individuals interact locally, the outcome is a stable plural patchwork of cross-cutting conflicts. By encouraging nonlocal interaction, digital media drive an alignment of conflicts along partisan lines, thus effacing the counterbalancing effects of local heterogeneity. The result is polarization, even if individual interaction leads to convergence. The model thus suggests that digital media polarize through partisan sorting, creating a maelstrom in which more and more identities, beliefs, and cultural preferences become drawn into an all-encompassing societal division.
SBIRs
9:00 kernel methods discussion with Rukan. Need to look at a way of using SD to look at “outlier-ness” maybe SD of all points – SD of all other points
10:00 experiment logger review
Got GeoPandas installed!
An example for a Windows Python 3.7 install in the directory where the whl files are located would be (in order):
Got the user downloading working. Smooth, with no surprises!
Write a class for creating corpora. It will take the settings from the UI and generate files with the appropriate wrapping
New view that includes cluster exclusion True/False
Dig up the code that produces percentage meta info
SBIRs
Asked Loren and Rukan to keep their parts of the Overleaf up to date
9:15 standup
Port Panda3d visualizer. Mostly done, but I’m having problems with objects being multiply defined. Turns out the setup() method was being called twice – Once when the class was instanced, once when the visualizer called setup() directly. Fixed.
Start outlining update of AI & Weapons paper – nope
In the past decade, we have witnessed the rise of deep learning to dominate the field of artificial intelligence. Advances in artificial neural networks alongside corresponding advances in hardware accelerators with large memory capacity, together with the availability of large datasets enabled practitioners to train and deploy sophisticated neural network models that achieve state-of-the-art performance on tasks across several fields spanning computer vision, natural language processing, and reinforcement learning. However, as these neural networks become bigger, more complex, and more widely used, fundamental problems with current deep learning models become more apparent. State-of-the-art deep learning models are known to suffer from issues that range from poor robustness, inability to adapt to novel task settings, to requiring rigid and inflexible configuration assumptions. Collective behavior, commonly observed in nature, tends to produce systems that are robust, adaptable, and have less rigid assumptions about the environment configuration. Collective intelligence, as a field, studies the group intelligence that emerges from the interactions of many individuals. Within this field, ideas such as self-organization, emergent behavior, swarm optimization, and cellular automata were developed to model and explain complex systems. It is therefore natural to see these ideas incorporated into newer deep learning methods. In this review, we will provide a historical context of neural network research’s involvement with complex systems, and highlight several active areas in modern deep learning research that incorporate the principles of collective intelligence to advance its capabilities. We hope this review can serve as a bridge between the complex systems and deep learning communities.
Tasks
Chat with Dave tonight? – Need to send links to papers, OpenAI, Stable Diffusion thread, Paul Sharre’s books, etc
SBIRs
Fill out reimbursement forms – done
Travel to Chirp – tried in Concur. Hopeless mess
More RCSNN2
9:15 standup
Experiment logger meeting – done
Reached out to Dr. Giddings on format of white paper
GPT Agents
Change table_user to have user ID as unique, primary key and see if update into works right – done
Add checkboxes for optional user attributes (requires downloading users for tweets)
No meeting is scheduled, so write up a status report
We offer comprehensive evidence of preferences for ideological congruity when people engage with politicians, pundits, and news organizations on social media. Using 4 years of data (2016–2019) from a random sample of 1.5 million Twitter users, we examine three behaviors studied separately to date: (i) following of in-group versus out-group elites, (ii) sharing in-group versus out-group information (retweeting), and (iii) commenting on the shared information (quote tweeting). We find that the majority of users (60%) do not follow any political elites. Those who do follow in-group elite accounts at much higher rates than out-group accounts (90 versus 10%), share information from in-group elites 13 times more frequently than from out-group elites, and often add negative comments to the shared out-group information. Conservatives are twice as likely as liberals to share in-group versus out-group content. These patterns are robust, emerge across issues and political elites, and exist regardless of users’ ideological extremity.
Tasks
Jim Donnies (winterize and generator) – done
BGE
ProServ
SBIRs
9:30 RCSNN design discussion – done
2:00 Meeting with Loren – done
Write up some sort of trip report
Reach out to folks from conference – done
Start on distributed data dictionary? Kind of?
Book
Roll in Brenda’s Changes – continuing
Ping Ryan for chapter/paper/article on authoritarians and sociotechnical systems
GPT Agents
Add cluster ID to console text when a node is clicked and a button to “exclude topic” that adds an entry to “table_exclude” that has experiment_id, keyword (or “all_keywords”), and cluster_id. These clusters are excluded when a corpora is generated.
Re-clustering will cause these rows to be deleted from the table
Add training corpora generation with checkboxes for meta-wrappers and dropdown for “before” or “after”
viztales 
You must be logged in to post a comment.