Chinese AI firm DeepSeek unveiled its GRM evaluation framework on 12 April 2025, aiming to refine generative models through meta-reinforcement learning. ❤️ #AIdevelopment #ChinaTechPolicy #CodingBenchmarks #generativemodels #MetaLlama4 #redrobot
#GenerativeModels
Do you want to apply deep learning-based methods to your own data and image analysis problems? Then #EMBLDeepLearning is for you!
📣 Apply by 25 November 2024 📣
➡️ https://s.embl.org/mac25-01
🗓️ 17 – 21 February 2025
📍 EMBL Heidelberg
Prerequisites for this workshop are programming experience in Python, including solid knowledge of operations with images, and first experience of applying deep learning algorithms.
I was waiting for this point:
New research paper explores the risk of generative models creating "Model Autophagy Disorder" (MAD) when training on their own synthetic data. Without fresh real data, future models may see a decline in quality and diversity, leading to amplified artifacts. Read the full study for insights on how to mitigate these effects. #AI #MachineLearning #GenerativeModels #DataScience #Research #Tech
Link to the paper: https://arxiv.org/pdf/2307.01850
Generative AI for databases
This new tool offers an easier way for people to analyze complex tabular data
* easy-to-use tool enables complicated statistical analyses on tabular data using just a few keystrokes
* combines probabilistic AI models w. programming language SQL
* provides faster/more accurate results
GenSQL: A Probabilistic Programming System for Querying Generative Models of Database Tables
https://dl.acm.org/doi/10.1145/3656409
#SQL #ML #GenerativeModels #GenSQL #DataScience #MIT #databases
Regarding that space of people that seem to seriously think that #generativeModels are coming for software developers in a big way… what do you think happens when the model bugs in the work? You don’t have software developers who spend the time architecting the system and understanding its behavior.
Companies using these „AI“ tools will be sitting there bleeding money while the human developers get started on even understanding the design in the first place.
#Modelling #animal #network data in #R using #STRAND
#OpenAccess
#JournalAnimalEcology
"the STRAND R package provides a suite of #GenerativeModels for #Bayesian analysis of animal social network data that can be implemented using simple R syntax. A #tutorial demonstrates how STRAND can be used to model #proportion, #count or #binary network data using #StochasticBlockModels and/or #SocialRelationModels"
https://besjournals.onlinelibrary.wiley.com/doi/full/10.1111/1365-2656.14021
So, impression about #generativeModels of the day. These technologies destroy human lives.
- they destroy our employment opportunities
- they destroy our agency, by confabulating quotes and creating entire personalities for stalkers to interact with
- their data centers take up so much electricity it displaces physical businesses that employ us, make our food and local services
- their data centers use up the water we need to drink
What about this entire design space is *not* hostile to us
Making the internet more hostile to human lives through #algorithms, #metrics and #generativeModels should be treated as an attack upon human rights and be responded to as such.
late realization from a dude who has not read theory, but belike - we should probably call out #LLMs and other #GenerativeModels as the tools of #classwar they are.
And that this is the *main* motivation why they basically get pushed into deployment at a point where they are unsafe to the people who interact with them, and can’t do any job to human-like qualities.
They exist to manufacture pretense of human replaceability.
Discover the latest AI tool, Nightshade, for protecting digital art, created by researchers at Chicago University. Learn more here
#AIArtProtection #BenZhao #digitalart #FutureAIDevelo #GenerativeModels #GlazeProject #zugtimes
https://www.zugtimes.com/are-the-days-of-generative-ai-numbered/
All right ... we're there now.
OK.. ormai ci siamo
I'm excited to see recent developments with LLMs may provide effective modeling of our aggregate values (multi-level, fine-grained), increasingly coherent over increasing context of meaning.
LLMs such as ChatGPT and Gemini already impress us with their ability to produce coherent (but not necessarily correct) abstractions from massive collections of data relevant to our interests.
Now, what if we run multiple instances of LLMs with diverse priors, collaboratively (at lower levels adversarial, at higher levels cooperative) selecting for increasing coherence over increasing context in the domain of present but evolving values?
Of the two orthogonal dimensions producing the expanding space of moral (right in principle) agency —our values, and our methods for their promotion — we already have much attention on our evolving methods (science, technology) and it appears we are on the cusp of rapid development of an effective model of our evolving values!
Cause for (Promethean) hope!
Research remains to be performed on theoretical and practical questions involving expected signal-to-noise improvements of "synthetic data" extracted from latent information in our datasets, regression due to synthetic data added to the data set, biases, gaming of the system, and on and on… Interesting, exciting, and dangerous.
#LLM #LargeLanguageModels #AI #GenerativeModels #SyntheticData #DecisionMaking #CollectiveDecisionMaking #Ethics #MetaEthics #Morality #ArrowOfMorality
We should introduce the securities that #UBI is supposed to give people, for many reasons.
But not because #generativeModels "#AI" is capable of replacing people. That‘s buying into destructive #critihype about the capabilities of the technology.
GenMods *cannot* replace people to the same quality, and they‘re giant tools to plaster shut the human creative sharing space with so much useless signal it drowns out human creativity.
started from @bobdoto
https://pkm.social/@bobdoto/111905286616083031
We do a lot of creative and interesting work here at @desy, including using #machinelearning #generativemodels to make particle collision simulations. Here's doctoral student Moritz Scham explaining his work on #DeepTreeGAN, essentially a "#DALL-E for #particles" - check out more on group's instagram: https://www.instagram.com/p/C12MbLtArUA/
FreeU: Revolution in AI generation – quality without additional training!
#FreeU #AI #GenerativeModels #ProbabilisticDiffusion #UNet #Deinterference #ImageGeneration #TextToVideo #Research #Innovation
#Perspectives #GenerativeModels Text-to-CAD: Risks and Opportunities https://thegradient.pub/text-to-cad/?utm_source=dlvr.it&utm_medium=mastodon
The discussion around the "privilege of art" that currently go around due to #GenerativeModels aka "#generativeAI" feel to me more and more like a roaring admission that we've created a society hostile to people having the time and space and livelyhood security to explore and develop their own artistic expression. And in good tech-solvism manner, instead people now propose to patch that over with fifteen-minute #consumerism by megacorp suppliers.
#Enoch from #StarWars #Ashoka seems like one of those creepy images created by ML #GenerativeModels.
I love it.
Bayesian Flow Networks
Alex Graves (https://en.wikipedia.org/wiki/Alex_Graves_(computer_scientist)) et al.
https://arxiv.org/abs/2308.07037
reddit/ML: https://old.reddit.com/r/MachineLearning/comments/15rrljw/bayesian_flow_networks
* parameters of indep. distributions modified w. Bayesian inference in light of noisy data, passed as input to NN that output 2nd, interdependent distribution
* generative procedure similar to reverse process of diffusion models h/e no forward process req'd
#MLarchitecture #MachineLearning #NeuralNetworks #BayesianFlowNetworks #Bayes #AlexGraves #GenerativeModels #BFN
Client Info
Server: https://mastodon.social
Version: 2025.04
Repository: https://github.com/cyevgeniy/lmst