Peakey Enterprise LLC

11/25/2024

💡 Innovation in Action: A Success Story with CUEs! 🤖

We’re thrilled to share how our partnership with CUEs, a leader in inspection robotics, took their technology to the next level!

👉 The Challenge: Automate the time-consuming process of completing inspection forms and modernize their solution.

👉 Our Solution:
Designed an AI-powered automation to save time and reduce errors.
Modernized their web and mobile applications for a smoother user experience.
Optimized their systems to boost efficiency and reliability.

🌟 The Results: Together, we delivered smarter, faster, and more intuitive solutions that not only met but exceeded expectations—helping CUEs continue to lead the way in their field.

Curious to learn how we made it happen? Read the full story here: https://www.peakey.com/customer-success-stories/

✨ When innovation meets collaboration, amazing things happen!

10/25/2024

Discover ANSWER, the next-generation AI solution for robotics, offering intelligent automation, cloud services, and custom skills management for industries like manufacturing, logistics, and healthcare. Enhance productivity with our easy-to-deploy, advanced AI-powered robots.

10/23/2024

🤖 We’re thrilled with ANSWER, but we want to hear from YOU!

👉 How do you think AI will change the robotics landscape in the next 5 years?

👉 What’s the biggest problem in robotics that AI could help solve?

Join the conversation in the comments! Let’s discuss the future of robotics together.

Drop your thoughts and share this post to spread the word!

https://www.peakey.com/answer/

Discover ANSWER, the next-generation AI solution for robotics, offering intelligent automation, cloud services, and custom skills management for industries like manufacturing, logistics, and healthcare. Enhance productivity with our easy-to-deploy, advanced AI-powered robots.

10/14/2024

🚨 Big News Alert! 🚨

💬 We’d love your help! Please like, share, and spread the word about ANSWER if you’re as excited about the future of robotics as we are! 🙌

Introducing ANSWER – the future of AI-powered robotics is here, and we're on a mission to make advanced robotics more accessible than ever before! ANSWER powered robot can understand your requests and complete them like a human would.🤖✨

Whether you're a tech enthusiast, innovator, or business leader, ANSWER is designed to empower YOU. Imagine building your next project with cutting-edge AI capabilities at your fingertips. 🔧💡

🎥 Check out our demo video below to see ANSWER in action! Trust us, you don’t want to miss this one!

📣 We're just getting started, and we want YOU to join us on this journey. Be among the first to get updates, exclusive insights, and sneak peeks. 📬

👇 Let’s build a community of future-thinkers together! 🙌 Get exclusive updates, behind-the-scenes insights, and be the first to know about our Kickstarter launch by signing up for our newsletter!

💌 Sign up here: https://mailchi.mp/peakey/answer-advanced-neurobotics-solutions-workflows-execution-and-robotics

🌟 Together, let's shape the next era of robotics! 🌟

01/03/2022

Wanted to share a quick example of our capabilities with Natural Language Processing (NLP). In this example, you can speak English in, hear the French translation back and see the user's sentiment (positive or negative). The specific models we used in this example were wav2vec2, distilbert, t5, and tacotron2 with mb_melgan.

To accomplish this we chained 4 machine learning tasks together.

1. Convert the audio to text.
2. Sentiment analysis (determine if the audio is positive or negative).
3. Convert the text from English to French.
4. Produce French spoken audio data.

Visit the link below to view the results.
https://buff.ly/3mMHtbN

08/18/2021

Studies have shown, nearly $1 trillion a year is lost to machine failures globally. Early and accurate prediction of degradation and failure allow for preventative maintenance measures to be employed to help reduce or eliminate machine failure and unplanned downtime.

We’ve recently released a whitepaper introducing our (patent pending) novel approach to predicting degradation and failure, demonstrate how our novel solution is able to produce results in a self-supervised solution, and achieve greater accuracy and lead time than was previously possible.

Predicting the Unpredictable: Implementation of an Advanced Machine Learning Method to Predict Degradation and Failure using Multivariate Time-Series Forecasting

Request your free copy in the comments below or visit https://www.peakey.com/whitepaper/predicting-the-unpredictable-adv-machine-learning/

06/23/2021

Peakey is proud to be named for a second year in a top B2B companies by Clutch!

Clutch is a B2B ratings and reviews platform where businesses from across the world can connect with service providers for their next big project. Market presence, project history, and verified client reviews are used to award winners. Former clients are interviewed over the phone and provide honest feedback that is published onto the company’s Clutch profile.

Thank you Clutch for being an awesome partner and for the award!

View our clutch profile: https://buff.ly/3vP2Vyn
Clutch press release: https://buff.ly/35HpCKb

05/26/2021

We're excited to announce that the machine learning software we've built will be used during the Indy 500 this year! The solution ingests real-time information from cars and predicts failures and deterioration giving the teams more time to plan pits, help increase speed, and support driver safety.

04/01/2021

Common sense in AI systems? Is it possible? Maybe not now, but as a community of researchers we are getting closer. At least that is the premise of the most recent article released by the chief AI scientist at Facebook Yann LeCun. https://buff.ly/3bDz3y9. The main idea in this article is that self-supervised learning produces models that have learned a better understanding of the nature of the task regardless of domain (computer vision, natural language processing, etc.). For example, in traditional computer vision supervised learning we tell the model this is an image of a dog, for all images of dogs we have. Then, when we ask it what a new image is, maybe it is correct in identifying the dog and maybe it isn't. However, if we apply the same technique after self-supervised learning we get a model that is correct more often and with less data needed. This leads to the conclusion that further research in self-supervised learning will further approximate common sense. We have done our own research of self-supervised learning with the framework lightly. . And we look forward to continuing our own research in this area, and passing the benefits on to our customers. The attached image shows a SSL trained model recognizing that these are similar without classes.

02/22/2021

We're hiring! Looking for a software developer who is passionate about building great things with new technologies.

https://www.indeed.com/job/software-developer-c598ec31f30dcce8

02/18/2021

Our latest efforts into cutting edge ml/AI, technologies has bore "fruit". The tool can take simple requests such as "a photo of an orange with peakey written on it" and you'll get an image of just that. Attached is the result of this request. It took several hours on the test machine to generate this image.

This expirement used big-sleep https://buff.ly/2NE3ieY. A model that combines CLIP https://buff.ly/3ozXkJp with bigGAN https://buff.ly/3bwTM5J to allow for text to image generation.

If this doesn't impress you I suggest taking a look at dall-e which is the same process but with the full power of GPT3 https://buff.ly/37ohcso . Even though this technology is in its infancy we are excited for it's possible use cases for automation in art, marketing, and other creative fields.

12/29/2020

In our latest research experiments we have delved into the area of
machine learning known as reinforcement learning (RL). The main
application of reinforcement learning is robotics, but there are also
other areas such as finance which may also take advantage of RL. In
this scenario, the agent/robot, is allowed to train in a simulated
environment that mimics the desired task in the real world. As you can
see in the image above, the humanoid figure is the agent who, after
millions of attempts, learns which actions to take to reach the red
ball. For those interested in what we used for our experiments here are
some links. OpenAI Gym is a toolkit for RL environments. https://buff.ly/2HruxBl .
pybullet-gym is an re-implementation of the 3D environments found in OpenAI Gym , but with PyBullet as the physics engine. https://buff.ly/3rxEDbr .
MuZero is a state of the art RL algorithm. https://buff.ly/3aTbl0M (edited)