Figure Eight Workflows Streamlines Multi-Step Automation for High-Quality AI Training Data

Creating high-quality data sets often requires completing a complicated maze of multi-step annotation jobs, making it difficult and expensive for companies to navigate setting up annotation tasks. Multi-step annotation projects can develop problems like reliance on expensive, scarce technical resources, latencies, manual error and unpredictable costs.

Society6 is home to hundreds of thousands of artists from around the globe. Artists upload more than 4,000 new creative assets every day, and Society6 must filter out low resolution images, as well as inappropriate or copyrighted material. Before turning to Figure Eight, Society6 would manually review its site every time it received a legal notice about any copyright or IP infringement.

Workflows makes it easier to break down complex annotation jobs into simpler jobs and connect them dynamically. For Society6, Workflows automates the process of separating out creative uploads into review buckets. This helps Society6 prioritize the review of inappropriate artwork and remove these assets at a faster rate. Society6 is a great example of how Workflows can perform multiple types of annotation jobs in the same workflow on the same initial data set without human intervention, as opposed to running different independent jobs and manually processing the intermediary data.

“Before Figure Eight, we had to manually scrub our site every time we received a legal notice. Legal notices come from artists, agencies, law firms, etc. Most of the notices coming through are for IP infringing content. Figure Eight helped us prioritize content and quickly make judgments on creatives that needed to be suspended,” said Jessica Mo, Associate Product Manager at Society6.

With Workflows, Society6’s internal artwork review team has been able to review almost 30,000 creative pieces in just two months, up from a few thousand reviewed pieces per month.

Marketing Technology News: Plannuh Raises $4 Million To Expand AI-Driven Marketing, Budgeting and Planning Platform

Workflows Lessens Annotator Burden, Reduces Costs

Workflows also provides the flexibility to target different contributors for every annotation step, so that only highly skilled contributors handle the most difficult steps.

“Overly complex data-annotation jobs increase the cognitive load on the global crowd tasked with labeling vast quantities of training data,” said Wilson Pang, CTO, Appen. “To help create high-quality machine learning data more effectively, we’ve developed technology that streamlines the annotation process. Workflows easily connects multiple, more specific jobs within large annotation projects to optimize the process for quality and improve the experience for both AI experts and the annotation crowd. By creating more granular annotation jobs, Workflows also delivers high-quality results faster, leading to fewer wasted resources and reduced costs when compared to large, complex annotation jobs.”

Workflows Featureset Makes Annotation Simple, Even in Complex Environments

Workflows offers customers like Society6 routing rules to control when data qualifies to be routed to the next step(s). Other benefits include:

• Plug-and-play operability: Workflows uses a graphical user interface to enable anyone, regardless of technical skill, to configure operators with routing rules.
• Automation: Workflows automates data routing and the aggregation of results into a single report, which eliminates bottlenecks of managing the annotation process manually.
• Improved quality: Workflows enables customers to break complex projects down into smaller tasks, which reduces cognitive load for contributors and optimizes for data-annotation quality.
• Lower costs at scale: Machine Learning-Assisted Data Labeling (MLADL) combines human annotation with machine learning to deliver annotated data up to 20 times faster at up to a 50% lower cost.

Marketing Technology News: Wealth Access Enhances User Productivity with DocuSign Integration