Blog

We’re proud to release the Sama-Coco dataset, a relabeling of Coco-2017. It’s free and ungated for any machine engineers to use in their projects. Happy Halloween!

This article is the first in a series regarding how Sama uniquely equips its clients with a data labeling process that complies with the emerging body of AI regulation to ultimately avoid liability risks and costly rework. NTD: worth going into the importance of Digital Trust?

Data curation tools help prioritize labels that will have the greatest impact on your models, quickly optimizing model accuracy while saving time & money.

What helps self-driving cars “see”? At their core, autonomous vehicle perception systems rely on object tracking algorithms to help understand their surroundings and ensure optimal safety.

What helps self-driving cars “see”? At their core, autonomous vehicle perception systems rely on object tracking algorithms to help understand their surroundings and ensure optimal safety.

Like many companies across the world, Sama is navigating a shifting economy with an eye for long-term sustainability.

Here are best practice guidelines for choosing the right tools and the right approach for developing a robust 3D quality rubric.

In this ebook, Sama answers data labeling FAQs such as “How much data do I need?”, “Which parts of my data should I annotate first?” and many more.

Sama is honored to have made the Inc. 5000 List of America’s Fastest-Growing Private Companies… for the third year in a row. Being included on the Inc. 5000 list is a direct reflection of the hard work, dedication, and resiliency of our talented teams.

In ML, 100% data quality isn’t necessary – it isn’t even feasible. Learn how to find the quality “sweet spot” to get your models performing as desired.

More shopping doesn’t need to lead to longer lines. Many retailers have implemented self-checkout lines to help shoppers get their errands done more quickly.

Sama is investing in ML-powered tools to optimize efficiency for your video annotation projects; introducing ML object tracking to increase throughput without sacrificing quality.

Sama is thrilled to announce new and improved features for 3D point cloud annotation to provide our clients with the most secure, cost-effective, and fastest path to quality 3D drivable area annotation at scale.

Many online shopping experiences end in disappointment. In 2020, $102 billion worth of products purchased online were returned to retailers — about 18% of all online sales.

Good annotation and testing practices are the foundations of building a great model. However, understanding what constitutes quality data is a tricky question.

Almost 15 years ago, Leila Janah founded Sama on the belief that ​​“talent is equally distributed, but opportunity is not.” Her vision — to give work, not aid — has since produced ripple effects throughout East Africa and the world.

Smart homes and appliances empower homeowners to remotely control and program appliances. They wouldn’t be so “smart” without high-quality labeled data.

Customers walk around stores, browse displays, walk down aisles, and stop to consider products. From the outside, it may not appear there is much to glean from these behaviors.

Today, Sama is thrilled to announce a host of new offerings which will allow us to continue to power our customers’ ML models with unmatched accuracy, scale, and efficiency.

In sports, AI can drive ticket and merchandise sales by studying purchase patterns from a large number of fans. AI is used in sports both on and off-pitch.

Here are some of the most frequently asked data labeling questions, along with recommendations for approaching your data annotation strategy holistically.

The purpose of a LiDAR annotation quality rubric is simple: it ensures that two people will score the same object in identical ways.

“Like so many immigrant children, I learned to believe in a dream that is as much American as it is universal: a dream of equal opportunity.” – Leila Janah

Continuous improvement and excellence are at the heart of everything we do, and we invite anyone, in and outside of the company, to help us on our journey.

Sama is committed to providing our workforce with professional development and upskilling opportunities, benefits, and a living wage.

Orbisk is focused on limiting the amount of food waste produced in restaurants, hotels, and cafes with their AI-powered food waste monitoring solution.

In-house data annotation can be expensive but can be helpful in early stages of ML production. Outsourcing data annotation is cheaper but security can be compromised.

A recent academic paper outlines how Sama uses Experiment-Driven Development to measure how improvements made to our platform increase annotation efficiency.

Swift partnered with Sama to cost-effectively scale data annotations for their CV model, without compromising the clinician-level accuracy they needed.

Sama is thrilled to partner with research center Mila, to help further the mission of creating innovative technologies to drive the AI industry forward.

Sama has been recognized in Fast Company’s list of the Next Big Things in Tech for our work with our partners at Orbisk.

Learn how Volumental partnered with Sama to accurately label the datasets that fuel the computer vision technology for their mobile foot scanning app.

Last week, Sama secured Series B, cementing the viability of purpose-driven companies—and hopefully inspiring others to pursue purpose and profitability.

Sama has raised $70M Series B funding to build the first end-to-end AI platform that enables teams to manage the complete AI lifecycle.

Find out how accurately labeled data is helping PolyPerception provide material recovery facilities with better visibility into their waste streams.

ML Assisted Annotation can help you generate high-quality pre-labeled and human-assisted annotations, for predictably higher quality data in half the time.

This whitepaper by Partnership on AI examines the working conditions of data enrichment professionals, and what can be done to improve them.

Project Guideline by Google partnered with Sama to help people who are blind run without a guide, using only a smartphone, headphones, and a yellow guideline.

Sama’s third annual Innovation Week is coming to a close, and once more, our teams have given us plenty to be excited about.

This week, researchers at MIT released a white paper evaluating Sama’s impact through a three-year Randomized Controlled Trial study. Here are their findings.

Tribe Dynamics helps customers get better ROI from influencer programs. Find out how partnering with Sama helped Tribe better serve clients and expand into new markets.

In this webinar, discover the challenges & solutions to 3D LiDAR annotation & 3D data sets for solving autonomous driving or driver assistance.

We asked experts working in the field about their thoughts on the role of humans in Machine Learning, and humans and the future of ML. 

How do you define training data quality and measure it? How do you improve it? We go into defining, measuring, and reviewing your training data quality.

87% of AI projects will never make it into production. Why? We asked ML experts.

Sama is an expert in efficiently designing annotation guidelines that enhance data quality. Gold Tasks refer to tasks that have been annotated perfectly.

In this series of three we’ll go into Experiment Driven Development and A/B Testing. EDD is fact-based development: based on evidence, not intuition.

Automated quality accelerators are technology innovations that are focused on reducing the amount of manual quality assurance time spent in QA processes.

In this series of three we’ll go into Experiment Driven Development and A/B Testing. EDD is fact-based development: based on evidence, not intuition.

In this series of three we’ll go into Experiment Driven Development. EDD is fact-based development: based on evidence, not intuition.

We reached out to various ML experts, asking them the questions: Why is high-quality training data so important? Why do so many projects fail in ML?

Introducing Factotum: an MIT-licensed, open source, kubernetes-oriented, general purpose docker container for devs/devops and custom CI/CD pipelines.

2021 Predictions: We asked a range of ML experts about what they believe will be the next big thing in AI and Machine Learning.

The Sama MLOps Pipeline: At Sama we decided to build our own automated training pipeline in order to limit costs, and to avoid tying ourselves to a particular cloud provider. 

Samasource is now Sama, the same team that powers the world’s most ambitious AI projects with high-quality and accurate data, but with a new name that represents our vision moving forward.

There’s no shortage of literature about ML, but we’ve compiled a list of 10 must-read books to add to your list!

In this article we summarize an approach that we have developed to speed up polygonal instance segmentation using machine learning.

Sama is thrilled to announce that Wendy Gonzalez is our new Chief Executive Officer.

Announcing our support for custom keypoint shapes in our training data platform trusted by the world’s leading AI teams, for vector image and video annotation.

Here’s a celebratory list of some of the women we look up to and have spearheaded development in AI and Machine Learning in 2020.

For the second consecutive year, Sama was a Terabyte partner of the McGill Engineering Hackathon, the largest annual hackathon run by the McGill Electrical, Computer, and Software Engineering Student Society.

We’re humbled and honored to announce that, as of today, Sama is the first AI company to receive the prestigious B Corp certification.

AI is often framed as something that’ll change our future, but many people aren’t aware of quite the extent to which AI currently used in society and everyday life.

Through our work with Vulcan, we have been awarded the winner of the 2020 AI Breakthrough Awards for the Best Image Processing Solution. 

Sama is honored on Inc. Magazine’s Annual List of America’s Fastest-Growing Private Companies—the Inc. 5000

From accidental Alexa purchasing to bias in recruitment, we have gathered 5 AI fails from the last few years.

We explain the various types of machine learning algorithms including supervised, unsupervised and reinforcement learning, as well as business use cases.

We’ve decided to take action both inside our organization and in partnership with local organizations to drive toward a more equitable society.

The growth of popularity in AI has been mirrored by a growing number of concerns surrounding privacy, security and ethical use of data.

We’re excited to introduce Chloe, an automated public chatbot service to support in the fight against the coronavirus.

Fast Company has recognized Sama as a finalist in the AI and Data category as part of their 2020 World Changing Ideas Awards.

The traffic light problem for autonomous vehicles is critical for all vehicle safety, and unlike human-drivers, AVs rely solely on computer vision systems to navigate the world around us.

Creating lasting social impact, understanding our carbon footprint and being environmentally sustainable are just few of our goals for sustainability and impact.

Sama video and 3D object tracking with frame-level labeling assists companies in quickly building models that better reflect real-world behavior.

Sama’s commitment to an ethical AI supply chain enables us to close the gender pay gap, by paying a living wage to all of our employees. 

As AI adoption expands, untapped communities are finding work at the cutting edge of AI. Here are 4 ways AI makes a positive impact on communities in East Africa.

During the REWORK Deep Learning Summit, Sama shared how top organizations obtain secure, high-quality training data, fighting AI bias in the process.

In this interview, we chat with Head of AI at Sama about AI trends to expect in 2020, as well as frequently asked questions about AI and machine learning.

Sama was proud to sponsor CodeJam 2019, an annual hackathon at McGill University, from November 15 – 17, 2019.

When Walmart set out to improve e-commerce search relevance, it needed a labeling partner who could scale with demand without sacrificing quality.

When we work with companies that want to use AI in e-commerce, we notice a few common barriers in AI adoption. Here are 4 reasons e-commerce needs AI to stay ahead.

ICCV 2019 provided a welcoming platform for the distribution and discussion of scholarly and technical work in computer vision.

From improved shopping experiences to increased buyer’s confidence, here’s what Nike’s use of AR in e-commerce means for the retail industry.

This list of computer vision insights shares how artificial intelligence is learning to understand and relate to the intensely visual world around us.

Here are a few things to consider, to strengthen your data security practices.

Here are 6 TED talks on AI Ethics anyone working in artificial intelligence should watch.

Data isn’t the only thing holding back artificial intelligence. Read more about some of the challenges and trends in AI.

Sama’s revamped toolset for 2D image vector segmentation is ideal for computer vision projects using vector shapes to structure training data.

Most media coverage discuss bias, fairness, and ethical use of AI, but the humanitarian aspect of the AI supply chain is often overlooked.

From facial recognition to AR/VR, computer vision is changing the way we interact with the world around us. Here are some highlights from CVPR 2019.

We’ve collected 6 computer vision infographics on everything from visual search to the history of computer vision from around the web.

Here are 7 organizations attending CVPR 2019 who are leading the way in computer vision, plus 3 noteworthy companies from around the web.

The Sama Hackathon in Costa Rica encouraged 56 students to look for ways to use data science to help reduce poverty and income inequality.

13 open source datasets for machine learning, including one dataset featured in the Fine-Grained Visual Categorization (FGVC) workshop at CVPR 2019.

Kirk Boydston, Training Data Specialist at Sama shares five considerations to move your machine learning model toward level 4 autonomous driving.

AI-enabled products come with their share of challenges. Here’s how Vulcan partnered with Sama to use artificial intelligence for wildlife conservation.

Sama exhibited and presented at the venerable 2019 Embedded Vision Summit in Santa Clara, California.

As a social business, we often face the dual challenge of generating social impact while balancing business needs. Learn how we’re measuring impact.

During Embedded Vision Summit 2019, Audrey Boguchwal will share four training data strategies that help AI teams avoid training data bias.

Today, we’re announcing the production availability of our new 3D annotation engine for the Sama.

To strike the balance between the latest insights from respected industry leaders and practical machine learning tactics to apply to real use cases, here’s our list of AI events to attend.

Today, Sama announces the availability of our latest image annotation toolset for advanced video object tracking.

In this post, we’ll present a simple overview of machine learning and how it helps computers solve complex problems.

ISO certification is a clear indicator that products and services meet the expectations of customers and that your data partner is qualified.

We’ll showcase how Sama’s web research and data cleaning services help create training data for Quid to build their NLP-powered data platform.

Data collection is a systematic strategy for gathering and measuring information from a variety of sources to get an accurate picture about a specific area of interest.

What is LiDAR and why is it such a hot topic? What are some of the key challenges around point cloud annotation? Find out in this post.

Last week, Sama visited the Auto.AI event, which bills itself as the platform bringing together the stakeholders who play an active role in the deep driving, computer vision, and sensor fusion.

With increased buzz around synthetic data, it is important to understand the advantages and limitations of this solution, and the overall affect on the application.

Synthetic data is system-generated data that mimics real data in terms of essential parameters set by the user.

The best way for a computer to gain knowledge is to start by showing it exactly what it is you want it to do. For this, we use training data.

We spotlight one of our clients, Markable, to show how our image annotation work can be used in ML to train image recognition algorithms.

Sama’s mission is to connect low-income people to digital work. This post explains the specifics of how this is done.

While certain applications of ML are prevalent in the media, there are lesser-known applications that are similarly revolutionary

Termed “planogramming,” visual merchandising is key in retail stores. The best stores find a balance between exciting customers without overwhelming them with deals shouting from every corner.

Here are five key tech and social impact trends from the 2016 World Economic Forum, starting with: It’s Not Enough to Do Less Bad.