• Thu. Apr 25th, 2024

Andrew Ng and Landing AI find to democratize AI for all company dimensions, push wider marketplace adoption


Feb 22, 2023
Andrew Ng and Landing AI seek to democratize AI for all company sizes, drive wider industry adoption


Check out out all the on-demand from customers classes from the Smart Security Summit here.

Andrew Ng’s cloud-primarily based platform for laptop eyesight, Landing AI, is taking on the arrival of artificial intelligence (AI) development amid businesses of all dimensions with its newest providing, LandingLens. The remedy promises to facilitate swift development and testing of pc vision AI jobs, without having the require for intricate programming techniques or prior AI working experience. 

“We began by discovering the producing sector, a single of the hardest industries in which to deploy computer system vision. Then we found the tools we had designed for producing, with comparatively number of modifications, can also be handy for a lot of other personal computer eyesight programs,” reported Ng, observed AI educational, and founder and CEO of Landing AI. 

The business announced nowadays that its flagship laptop or computer eyesight solution, LandingLens, is now out there for a no cost trial, coupled with a new pricing scheme that allows spend-as-you-go usage over and above the first demo period.

“With the new system, we aim to extend our tool’s use instances across many other industries,” Ng told VentureBeat. “To me, it is about reaching our objective of democratizing the creation of AI.”


Clever Stability Summit On-Demand from customers

Find out the essential job of AI & ML in cybersecurity and industry unique situation studies. Observe on-demand from customers periods these days.

Watch Here

“We want absolutely everyone to start for totally free and test it out to realize its use scenarios. We’re keen to make it offered for far more folks,” he claimed. 

A knowledge-centric approach to AI and laptop or computer vision

According to Ng, the platform’s data-centric AI technique focuses on data as an alternative of code, and as various industries progressively embrace AI solutions, a essential change is needed to unlock the complete potential of this technological know-how. 

LandingLens prioritizes boosting data good quality for AI versions, thereby enabling its features, even in instances exactly where businesses have confined data readily available for instruction the AI products, a widespread obstacle encountered by most companies. The “data-centric” tactic consists of coaching AI models to purpose proficiently with modest quantities of quality facts alternatively than relying exclusively on the wide datasets that usually underpin AI apps in substantial-scale internet businesses.

“Over the very last number of yrs, we did a lot work with prospects that frequently experienced little datasets. In the course of these ordeals, we found several technology techniques and optimizations that now allow our algorithm to function nicely on smaller datasets,” stated Ng.

He discussed that the product was educated on a ResNet dataset for impression recognition, and in the backend, LandingLens’s pretrained algorithm makes use of AI-based mostly automated hyperparameter tuning, enabling it to do the job well with datasets of every dimension. When info is passed by means of the design, it’s optimized through quite a few methods to deliver very well-analyzed, high-quality output and in depth insights. 

Just lately, therapeutic antibody discovery firm OmniAb utilised LandingLens to correctly automate its visible inspection approach, considerably rising effectiveness and throughput. In addition, the platform aided OmniAb in increasing AI accessibility inside of its group for use situations that require folks who are not substantial-level scientists. 

How does it function? 

To preserve facts consistency inside of LandingLens, the platform utilizes an advanced labeling technological innovation that routinely detects and corrects mislabeled illustrations or photos, boosting general info quality. 

This collaborative labeling solution will allow numerous consumers to label pictures and facilitates the system of achieving a consensus as a result of facts cloud and edge unit deployment abilities. As a consequence, deploying and screening your model can be reached with just a couple of clicks of the mouse. People can pick the deployment option that very best fits their requirements, ranging from a windows application to a programmatic API.

Furthermore, LandingLens employs a continuous-finding out system that ensures that the developed design remains up to date by integrating new info from the deployment ecosystem to retrain the product.

“We want to make the product growth workflow quick for buyers. The regular strategy to establishing AI products has generally been labeling, training to deployment. We want to relieve this development workflow by having people not publish a lot code, but aim a lot more on information entry,” extra Ng. 

Graphic resource: Landing AI.

Landing AI’s future focus on personal computer eyesight

Ng stated the enterprise would proceed to focus on producing the LandingLens platform as a one resource that serves several computer vision apps.

“Use circumstances in computer eyesight are now trying to keep us extremely occupied. Lots of clients throughout industries are requesting us to incorporate additional characteristics for conditions these types of as streamlining heterogeneous facts. So our latest roadmap consists of a large amount extra function to do in pc eyesight,” reported Ng. 

As a result of the LandingLens platform, Ng aims to clear up troubles located currently with customization or longtail AI design advancement, which he sees as the most major barrier to common AI adoption. 

“The only way for businesses to unlock maximum worth from their AI projects is when they have the liberty to customize their AI system as they will need. They can do this by engineering the info alternatively than the code. This way, companies can alter to the shifting current market specifications and establish greater styles utilizing lesser human means,” defined Ng. “So, I’m psyched about facilitating the aim of further more democratizing access to AI creation.”

The enterprise is pursuing programs in automotive, electronics and clinical product producing sectors. Ng stated embracing a information-centric AI methodology and utilizing AI and deep finding out-based solutions for laptop or computer eyesight situations will reward this varied vary of industries.

VentureBeat’s mission is to be a electronic town square for technological decision-makers to get information about transformative enterprise technology and transact. Discover our Briefings.

Leave a Reply

Your email address will not be published. Required fields are marked *