Sunday, August 14, 2022
HomeBig DataReport: 37% of ML leaders say they do not have the info...

Report: 37% of ML leaders say they do not have the info wanted to enhance mannequin efficiency

We’re excited to carry Rework 2022 again in-person July 19 and just about July 20 – 28. Be part of AI and information leaders for insightful talks and thrilling networking alternatives. Register immediately!

A brand new report by Scale AI uncovers what’s working and what’s not working with AI implementation, and the perfect practices for ML groups to maneuver from simply testing to real-world deployment. The report explores each stage of the ML lifecycle – from information assortment and annotation to mannequin improvement, deployment, and monitoring – to be able to perceive the place AI innovation is being bottlenecked, the place breakdowns happen, and what approaches are serving to firms discover success.

The report’s purpose is to proceed to make clear the realities of what it takes to unlock the complete potential of AI for each enterprise and assist empower organizations and ML practitioners to clear their present hurdles, be taught and implement finest practices, and in the end use AI as a strategic benefit.

For ML practitioners, information high quality is without doubt one of the most essential components of their success, and in line with respondents, it’s additionally essentially the most tough problem to beat. On this examine, greater than one-third (37%) of all respondents mentioned they don’t have the number of information they should enhance mannequin efficiency. Not solely do they not have number of information, however high quality can also be a difficulty — solely 9% of respondents indicated their coaching information is free from noise, bias and gaps. 

The vast majority of respondents have issues with their coaching information. The highest three points are information noise (67%), information bias (47%) and area gaps (47%).

Most groups, no matter business or degree of AI development, face comparable challenges with information high quality and selection. Scale’s information means that working intently with annotation companions can assist ML groups overcome challenges in information curation and annotation high quality, accelerating mannequin deployment. ML groups that aren’t in any respect engaged with annotation companions are the most probably to take better than three months to get annotated information. 

This survey was performed on-line inside the USA by Scale AI from March 31, 2022, to April 12, 2022. Greater than 1,300 ML practitioners together with these from Meta, Amazon, Spotify and extra had been surveyed for the report.

Learn the full report by Scale AI.

VentureBeat’s mission is to be a digital city sq. for technical decision-makers to realize data about transformative enterprise know-how and transact. Be taught extra about membership.



Please enter your comment!
Please enter your name here

Most Popular

Recent Comments