Walker Lee

Cofounder of @DataEngines, entrepreneur, management consultant, iaidoka, musician(?), father, etc. I'd rather ask questions than pretend I have all the answers.

[Walker Lee @WalkerLee]

@RSnakeCableCo's Ultimate Tone Giveaway - $1265 Prize Package
@Win_Gear_Here @giveawaywinfree @winmusicstuff @ContestWins @SweepsAdvantage
#giveaway https://t.co/HpqtUORx3U

[Walker Lee @WalkerLee]

Win @RSnakeCableCo - 2x 15' Custom Cables / MXR Custom Shop Hybrid Fuzz Giveaway #giveaway #fuzz #OD https://t.co/AtPcWMKGM3

[Walker Lee @WalkerLee]

@RSnakeCableCo's Ultimate Tone Giveaway - $1054 Prize Package
@Win_Gear_Here @giveawaywinfree @winmusicstuff @ContestWins @SweepsAdvantage
#giveaway https://t.co/k3NbgthEOy

[Walker Lee @WalkerLee]

@RSnakeCableCo's Ultimate Tone Giveaway - $1054 Prize Package
@Win_Gear_Here @giveawaywinfree @winmusicstuff @ContestWins @SweepsAdvantage
#giveaway https://t.co/k3NbgthEOy

[Walker Lee @WalkerLee]

Win @RSnakeCableCo - 2x 15' Custom Cables / @EarthQuakerDev Flexi Loops Giveaway #giveaway https://t.co/S23QfAccZi

[Data Engines @dataengines]

@rajiinio Readers interested in unsupervised evaluation upon deployment - a fundamental problem in #aisafety - check out our recently released Python package, ntqr, detailing how logic and algebra alone can help us here.
https://t.co/vbpncIoZRw

[Data Engines @dataengines]

@WIREDScience , @nytimes how many experts in #aisafety know there is an exact, algebraic solution to the problem of error independent binary classifiers being tested on unlabeled data? Algebraic numbers can protect us from hallucinating LLMs.
https://t.co/vbpncIoZRw

[Data Engines @dataengines]

The latest release (v0.1.5) of the ntqr Python package is out - building out the logic of evaluation in unsupervised settings so we can have provably safe evaluations of noisy agents when we give them tests for which we have no answer keys!
https://t.co/vbpncIoZRw

[Data Engines @dataengines]

Any intelligent being, whether human or robotic, would benefit from understanding the logic of evaluation in unsupervised settings to protect itself from its own mistakes. Check out how we are building it,
https://t.co/vbpncIoZRw

[Data Engines @dataengines]

If you believe, like @steveom and @tegmark , that we should have provably safe AI, check out the logic of evaluation in unsupervised settings that we have been building since 2010 with our first patent.
https://t.co/vbpncIoZRw

[Data Engines @dataengines]

New logics are being discovered. 🧵 1/n
https://t.co/aAsHL8PL2s

[Data Engines @dataengines]

Another technique to add is the use of logic and algebra to evaluate noisy AI in unsupervised settings, like so 👇
https://t.co/vbpncIoZRw

[Data Engines @dataengines]

If you are interested in how logic and algebra can help you build safer AI, check out the #jupyternotebooks in the tutorials for our NTQR Python package 👇
https://t.co/vbpncIoZRw

[Data Engines @dataengines]

Check out the Jupyter notebooks in the documentation for the NTQR Python package that explain the logic of evaluation using unlabeled data - a fundamental problem in #AISafety
https://t.co/vbpncIoZRw
#MachineLearning #Evaluation #Python #JupyterNotebooks

[Data Engines @dataengines]

@WIREDScience When are you going to alert your readers to the existence of a logic for evaluation of noisy AI algorithms that can make us safer?
https://t.co/vbpncIoZRw

[Data Engines @dataengines]

The newly released Python package, ntqr, by our science lead, @andrescorrada , contains Jupyter notebook tutorials explaining the logic and algebra of evaluation using unlabeled data.
https://t.co/vbpncIoZRw

[Data Engines @dataengines]

@xriskology Trust no one is our scientific working motto. True safety is giving the users the power monitor their own AI - grade a gaggle of noisy algorithms even when you do not have the answer keys to the tests you give them. 👇
https://t.co/q0xhIUgsq0

[Data Engines @dataengines]

Most importantly - black box algorithms are enough to create a complete logic of evaluation for noisy AI agents. This ameliorates the principal/agent monitoring paradox in unsupervised evaluation. https://t.co/X5vXThjKDv

[Data Engines @dataengines]

If you are concerned about the safety of the AI you are developing, consider learning about the logic of evaluation on unlabeled data. #AISafety #evaluation #unlabeleddata #groundtruth #AI #machinelearning
https://t.co/q0xhIUgsq0

[Data Engines @dataengines]

Those interested in discovering the logic of evaluation on unlabeled data with Python, consider 👇
https://t.co/q0xhIUgsq0