About
Articles by Andrés
-
Engineering Nearly Error-Independent LLMs: Lessons learned
Engineering Nearly Error-Independent LLMs: Lessons learned
By Andrés Corrada-Emmanuel
-
Thoughts on the "Leading with AI Responsibly" Conference
Thoughts on the "Leading with AI Responsibly" Conference
By Andrés Corrada-Emmanuel
Contributions
Activity
-
The rapid adoption of AI into healthcare processes creates an urgent need to review existing regulatory frameworks. Stanford HAI convened a group of…
The rapid adoption of AI into healthcare processes creates an urgent need to review existing regulatory frameworks. Stanford HAI convened a group of…
Liked by Andrés Corrada-Emmanuel
-
Found this nice blog post by Faraday on how they handle bias in their AI predictions - https://lnkd.in/e6GSNSzv
Found this nice blog post by Faraday on how they handle bias in their AI predictions - https://lnkd.in/e6GSNSzv
Shared by Andrés Corrada-Emmanuel
-
Been having some fun playing with Physics Influenced Machine Learning (PIML). Specifically, training Neural Nets to approximate incompressible flows…
Been having some fun playing with Physics Influenced Machine Learning (PIML). Specifically, training Neural Nets to approximate incompressible flows…
Shared by Andrés Corrada-Emmanuel
Experience & Education
-
Data Engines Corporation
View Andrés’s full experience
See their title, tenure and more.
or
Patents
-
System and Method For Imaging Through An Irregular Water Surface
Issued US 7630077
A way to use the polarization of light to create a fully submerged periscope that can see a clear view of the scene above by unscrambling the real time shape of the water. An example of using physics and computers to create a new kind of camera.
Other inventors -
-
Sequential, nonparametric speech recognition and speaker identification
Issued US 6029124
This speaker identification system won 6 out of 9 testing conditions in a NIST 1998 speaker identification evaluation of state of the art systems.
Other inventors -
-
Measuring Web Browser Tag Properties Without True Unique Tags
Filed US 14/492,332
The methodology of ground truth inference (figuring out how accurate your detectors are when no ground truth is available) applied to a problem in advertising tech - figuring out how accurate your unique tags are when you don't know the true unique tags. Notable in this patent is the idea that the accuracy can be measured with just two detectors (BlueCava IDs and browser cookies) because they have different types of error.
Other inventors -
-
METHOD FOR INFERRING ATTRIBUTES OF A DATA SET AND RECOGNIZERS USED THEREON
Filed US PCT/US2011/062772
From the patent application:
"The present invention generally relates to a method for inferring attributes of
a data set and/or recognizers used on the data set and, more specifically, to a method
for inferring a frequency of labels of the data set, a confidence of the labels, an accuracy
of the recognizers, an error rate of the recognizers and, in the case of a sequential data
set, the length of the data set, without knowledge of the correct labels to apply to the
data…From the patent application:
"The present invention generally relates to a method for inferring attributes of
a data set and/or recognizers used on the data set and, more specifically, to a method
for inferring a frequency of labels of the data set, a confidence of the labels, an accuracy
of the recognizers, an error rate of the recognizers and, in the case of a sequential data
set, the length of the data set, without knowledge of the correct labels to apply to the
data set."
Or informally - a method for measuring the accuracy of recognizers when there is no ground truth available for doing so!
Recommendations received
-
LinkedIn User
1 person has recommended Andrés
Join now to viewMore activity by Andrés
-
Richard Feynman called turbulence the most important unsolved problem in classical physics. Solving a major question related to vortices has also…
Richard Feynman called turbulence the most important unsolved problem in classical physics. Solving a major question related to vortices has also…
Liked by Andrés Corrada-Emmanuel
-
AI is great, but what about risks? Responsible AI came out with an early attempt at helping firms address some of the risks with their Top-20…
AI is great, but what about risks? Responsible AI came out with an early attempt at helping firms address some of the risks with their Top-20…
Liked by Andrés Corrada-Emmanuel
-
What a week in AI. (And it’s just Wednesday!) Here the plot of the trilogy: „A new hope“: Meta releases Llama 3.1 which outperforms ChatGPT 4o on…
What a week in AI. (And it’s just Wednesday!) Here the plot of the trilogy: „A new hope“: Meta releases Llama 3.1 which outperforms ChatGPT 4o on…
Liked by Andrés Corrada-Emmanuel
-
One of my biggest learnings from my career at Microsoft was that we should have open-sourced more of the work that my team did there. We had an…
One of my biggest learnings from my career at Microsoft was that we should have open-sourced more of the work that my team did there. We had an…
Liked by Andrés Corrada-Emmanuel
-
Alianza: nuvu y Blend impulsarán soluciones de Inteligencia Artificial Esta adquisición, única en su clase dado que ambas compañías tienen como core…
Alianza: nuvu y Blend impulsarán soluciones de Inteligencia Artificial Esta adquisición, única en su clase dado que ambas compañías tienen como core…
Liked by Andrés Corrada-Emmanuel
-
Do neural networks dream of internal goals? We confirm RNNs trained to play Sokoban with RL learn to plan. Our black-box analysis reveals novel…
Do neural networks dream of internal goals? We confirm RNNs trained to play Sokoban with RL learn to plan. Our black-box analysis reveals novel…
Liked by Andrés Corrada-Emmanuel
-
How can thoughtful use and regulation of healthcare AI help to protect patients and assist doctors? Stanford HAI Associate Director Curt Langlotz led…
How can thoughtful use and regulation of healthcare AI help to protect patients and assist doctors? Stanford HAI Associate Director Curt Langlotz led…
Liked by Andrés Corrada-Emmanuel
-
🎙️ New Episode Alert: Responsible AI In Action #21 🎙️ We're thrilled to welcome Cansu Canca, Ph.D., Founder+Director of AI Ethics Lab and Director…
🎙️ New Episode Alert: Responsible AI In Action #21 🎙️ We're thrilled to welcome Cansu Canca, Ph.D., Founder+Director of AI Ethics Lab and Director…
Liked by Andrés Corrada-Emmanuel
-
Do you remember at school when you had to include your workings to show how you got the answer? #AI might get the correct answer but its workings…
Do you remember at school when you had to include your workings to show how you got the answer? #AI might get the correct answer but its workings…
Liked by Andrés Corrada-Emmanuel
-
We're excited to launch our new AI + Health Fireside Chat Series, featuring leading experts discussing the future of #AI for #healthcare! Hosted by…
We're excited to launch our new AI + Health Fireside Chat Series, featuring leading experts discussing the future of #AI for #healthcare! Hosted by…
Liked by Andrés Corrada-Emmanuel
-
It's a wrap! Huge thanks to all our speakers and attendees for making the Vienna Alignment Workshop 2024 an amazing success! Stay tuned for videos…
It's a wrap! Huge thanks to all our speakers and attendees for making the Vienna Alignment Workshop 2024 an amazing success! Stay tuned for videos…
Liked by Andrés Corrada-Emmanuel
Other similar profiles
Explore collaborative articles
We’re unlocking community knowledge in a new way. Experts add insights directly into each article, started with the help of AI.
Explore More