Label Studio 1.12.0 🚀Automate & Evaluate Labeling Predictions Using LLMs & ML Models

Heartex is now HumanSignal!

Max Tkachenko

Co-founder, HumanSignal and Label Studio

We have some exciting news! In 2019, Michael, Nikolai, and I recognized one of the biggest challenges in machine learning was access to high-quality training data. This led us on an incredible journey building Label Studio, which has since become the most widely adopted open source data labeling platform, enabling more than 250,000 data scientists and annotators to label 200+ million pieces of data.

At the same time, we built a strong commercial business as Heartex, delivering our customers a fully managed Label Studio Enterprise platform, including advanced features for automation, team management, and improving data quality.

Now, we find ourselves ready to embark on the next stage of our journey, which also comes with a new identity: Heartex is now HumanSignal.

Why HumanSignal, you might ask? The answer lies at the core of our belief that the future of AI is not just about complex algorithms or massive computing power. It is about people. It is about the signal that humans provide, which powers these models, helping them to adapt, learn, and align with the needs of organizations and society at large.

In the last year, we’ve seen a sea change in the machine learning community in particular, and the tech community in general, with the rise of foundation models. These massive models require vast amounts of data to train them. Model developers have taken to scraping the internet for data to feed these models, using unsupervised learning to automate this daunting task. However, despite this "throw everything into the model" approach, these models still required retraining with human feedback to make them safe and valuable for broader release.

In this exciting new time of big models built from even bigger data, a critical component to reduce bias and increase accuracy is human feedback, or a human signal.

It’s our mission to help organizations take advantage of these foundation models and guide them to construct distinctly tailored models—while increasing accuracy and reducing bias—using high-quality data encoded with human feedback.

Label Studio—both the open source community edition and Enterprise edition—will continue to be at the core of our mission as we equip data science teams with the tools to build distinctly tailored AI models using proprietary data coupled with human feedback. We will focus on the quality and efficiency of human feedback, or signal, helping you take advantage of new foundation models and methodologies rapidly emerging in this space.

And don’t worry, our beloved Heidi the opossum will remain the official mascot and spirited cheerleader for the Label Studio Community!

You can read more about this change in Michael’s blog post on the HumanSignal website, and be sure to check out the new company homepage and identity. We’re excited about what’s next and look forward to continuing this journey with you. Everyone from HumanSignal wants to thank you for being part of the Label Studio community!

Related Content