The MLX is Apple's framework for machine learning applications on Apple silicon. The MLX examples repository provides a set of examples for using the MLX framework. This includes examples of:
✅ Text models such as transformer, Llama, Mistral, and Phi-2 models
✅ Image models such as Stable Diffusion
✅ Audio and speech recognition with OpenAI's Whisper
✅ Support for some Hugging Face models
@ramikrispin@BenjaminHan How do this and corenet (https://github.com/apple/corenet) fit together? The corenet repo has examples for inference with MLX for models trained with corenet; is that it, does MLX not have, e.g., activation and loss fns, optimizers, etc.?
@Lobrien@BenjaminHan The corenet is deep learning application where the MLX is array framework for high performance on Apple silicon. This mean that if you are using mac with M1-3 CPU it should perform better when using MLX on the backend (did not test it myself)
A former US military intelligence official released a letter on Monday that explained to his colleagues at the Defense Intelligence Agency (DIA) that his November resignation was in fact due to “moral injury” stemming from US support for Israel’s war in Gaza and the harm caused to Palestinians.
Harrison Mann, an army major, would be the first known DIA official to quit over US support to Israel.
Man said he felt shame and guilt for helping advance US policy that he said contributed to the mass killing of Palestinians.
“At some point, whatever the justification, you’re either advancing a policy that enables the mass starvation of children, or you’re not,” Mann wrote.
Version 0.12.0 of the skforecast Python library for time series forecasting with regression models was released this week. The release includes new features, updates for existing ones, and bug fixes. 🧵👇🏼
(2/2) Here are some of the new features:
✅ Ability to forecast multiple series with different lengths and/or different exogenous variables per series.
✅ Bayesian hyperparameter search is now available for all multiseries forecasters using optuna as the search engine.
✅ New forecasting models based on deep learning models (RNN and LSTM)
✅ New methods for creating prediction intervals
On June 15th, my colleague Mónica and I from @EA SEED will be presenting some of our work on #MachineLearning tools for #GameAudio at #AESEurope in Madrid. Really looking forward to visiting UPM again!
Looking for better data splits for #machinelearning? Look no further than astartes, a #pyOpenSci package from Jackson Burns, Kevin Spiekermann, and himaghna!
astartes is an #openscience, #opensource#Python package that implements many similarity- and distance-based algorithms to partition data into more challenging splits. Separate from astartes, you can use these splits to better assess out-of-sample performance with any ML model of choice.
Wie wird #KI & #MachineLearning die Software-Entwicklung beeinflussen? Diskutiert mit bei der uphillconf 2024, die wir als Bronzesponsor unterstützen. Es sind nur noch wenige Workshop-Tickets verfügbar! https://www.uphillconf.com/
Vraiment cool la série #Machine sur @arte 🔥
Un mélange explosif entre kungfu et lutte des classes avec Margot Bancilhon et la participation improbable de Joey Starr
Franchement à voir
Interpreting the LHC collisions is extremely data-intensive, and #CMSPaper 1282 describes how modern software techniques so our software (and #machinelearning) can run on many different platforms/processors and still efficiently and transparently reconstruct our collisions https://arxiv.org/abs/2402.15366
We’re so excited to announce the support of survival analysis for time-to-event data across tidymodels!
• The tidymodels framework is a collection of R packages for modeling and machine learning using tidyverse principles.
• Survival analysis is now a first-class citizen in tidymodels, giving censored regression modeling the same flexibility and ease as classification or regression.
@stefan Wait. I only get to see one at the time?! I mean, I build one, and then it show the machine working without any neighbours... How come there's neighbours in this video?
If you had the feeling that the online discussion about COVID-19 vaccines was biased depending on the actors, you are right.
Using #MachineLearning and #NetworkScience we have shown that being a human or a bot, verified or unverified (according to previous Twitter rules) and political leaning were relevant factors for choosing the words in posts and, accordingly, the corresponding emotions to trigger.
A genuine computational social science study, led by Anna Bertani for her Msc thesis, now published also in collaboration with Riccardo Gallotti and Pierluigi Sacco
the ask envision on the @letsenvision app is cool, I loaded a 50 page pdf and it just do a RAG on it and answers my questions comprehensively, people should use it more, I hope in future I can load entire folders of document on desktop to do rag #machineLearning#AI#blind#disability
Calling all data enthusiasts: ever heard of Orange (https://orangedatamining.com/)? Recently stumbled upon this tool for data mining and machine learning. It's Python-based and completely open-source. Sounds pretty good to me? Any users here?
Most of the Artificial Neural Net simulation research I have seen (say, at venues like NeurIPS) seems to take a very simple conceptual approach to analysis of simulation results - just treat everything as independent observations with fixed effects conditions, when it might be better conceptualised as random effects and repeated measures. Do other people think this? Does anyone have views on whether it would be worthwhile doing more complex analyses and whether the typical publication venues would accept those more complex analyses? Are there any guides to appropriate analyses for simulation results, e.g what to do with the results coming from multi-fold cross-validation (I presume the results are not independent across folds because they share cases).
@RossGayler
Aha, well yes it entirely depends on the question at hand and the experimental design. So eg. One major distinction is whether you are trying to say something about a model, a family of models, or the data. Parametric statistics is for inference over samples of a definable population, so eg. a point estimate of accuracy on held out data is fine if all youre trying to do is make a claim about a single model since there is no "population" you are sampling from. If youre trying to make a claim about a class of models then now you are sampling from the (usually) real valued, n-dimensional model space, so there the usual requirements for random sampling within parameter space would apply.
Making a claim about the data is much different, because now you have a joint analysis problem of "the effects of my model" and "the effects of the data" (neuroscientists love to treat the SVMs in their "decoding" analyses as neutral and skip that part, making claims about the data by comparing eg. Classification accuracies as if they were only dependent on the data. Even randomly sampling the subspace there doesnt get rid of that problem because different model architectures, training regimes, etc. Have different capacities for classifying different kinds of source data topologies, but I digress.)
For methods questions like this I try and steer clear of domain specific papers and go to the stats lit or even stats textbooks, because domain specific papers are translations of translations, and often have uh motivated reasoning. For example, the technique "representational similarity analysis" in neuro is wholly unfounded on any kind of mathematical or statistical proof or theory, and yet it flourishes because it sounds sorta ok and allows you to basically "choose your own adventure" to produce whatever result you want.
For k-fold, its a traditional repeated measures problem (depending on how you set it up). The benchmarking paradigm re: standard datasets and comparing accuracy is basically fine if the claim you are making is exactly "my model in particular is more accurate on this particular set of benchmarks." Youre right that even for that, to get some kind of aggregated accuracy you would want an MLM with dataset as random effect, but since the difference in datasets is often ill defined and as you say based in convenience im not sure how enlightening it would be.
Would need more information on the specific question you had in mind to recommend lit, and I am not a statistician I just get annoyed with lazy dogshit and think stats and topology (which is relevant bc many neuro problems devolve into estimating metric spaces) is interesting rather than a nuisance.
