Anari AI

23/12/2022

The Anari AI team wishes you a joyous holiday season!
Merry Christmas and Happy New Year! 🎉🎄✨

22/11/2022

We're looking forward to AWS re:Invent 2022 | Amazon Web Services 💥

Let's meet! always leaves room for new learning and discussion. 🙌

➡ Nov. 28 - Dec 2, 2022, Las Vegas, NV | Stojanovic

08/11/2022

MEET US TODAY at the iMerit ML DataOps Summit 🚀
iMerit Technology Services x TechCrunch

➡ Find our team at Anari AI virtual booth where you can learn more about the first "System-on-Cloud" - Thor X, and discuss the state of the ML DataOps landscape with Ksenija Ulićević, Jaksa Vucurevic, Mia Vujović, and Vasilije Pantić.

We're excited to be exhibitors and partners of the event that gathers 1,000+ attendees and 30+ leading AI and ML speakers from Cruise, Facebook AI, AI GE Life Sciences, Magic Leap, MIT, iMerit, and more.

*Registrations: https://tcrn.ch/3zWcMYm

25/10/2022

We’re happy to share that Anari AI will be a partner at iMerit ML DataOps Summit (iMerit Technology Services) organized in partnership with TechCrunch on November 8, 2022.

iMerit ML DataOps Summit will gather 1,000+ attendees and 30+ leading AI and ML speakers from Cruise, Facebook AI, AI GE Life Sciences, Magic Leap, MIT, iMerit, and more, all sharing their expertise on the current and future state of the ML DataOps landscape.

You can find us at the Anari AI virtual booth. 🚀

For more information and registration (use the Anari AI code) visit: https://tcrn.ch/3gEyg4S

03/10/2022

We're happy to be recognized as one of the most influential AI chip companies, among IBM, NVIDIA, Intel, Apple, and other industry giants!

🎂 This recognition comes on our 2nd birthday and we couldn't be more proud of what our team has achieved so far. 🎊

Cheers to many more years to come! 🥂

Top Artificial Intelligence (AI) Chip Companies Leading The Way In 2022. AI chips are a new generation of microprocessors created primarily to handle tasks requiring artificial intelligence more quickly and efficiently.

12/09/2022

Working with LiDAR or photogrammetry?

Thor X approaches both! ✅

Usually, point clouds are created by two main techniques, 3D scanning by LiDARs or photogrammetry, where each approach has its unique advantages and disadvantages. Different industries prefer one or the other, due to the point cloud density, precision, and other traits.

Thor X works with both LiDAR and photogrammetry-generated point clouds. Technologies used for both of these approaches are different, and so are the features they usually produce as an output, which inspired Anari’s team to optimize its first System-on-Cloud to cover all of the most relevant cases and features.

Find more details here: https://anari.ai/thor-x/

08/09/2022

We’re pleased to welcome Jovan Pavlov to the Data Annotator position on our team! We’re excited to have Jovan on our journey and look forward to empowering our team with such a great and motivated person!

Jovan, welcome to Anari AI! 🎉

30/08/2022

APPROACHING RAPIDLY GROWING DATA IN THE GEOSPATIAL INDUSTRY ⬇️

The geospatial industry is facing rapid data growth, which causes many problems from data storage to data processing and optimization of the process.

❗️The problem with processing and annotating hundreds of GB and TB of 3D point clouds is evident if we only reflect back to the example of manually annotating 3 billion points. For that amount of points, it would take 12 000 man-hours for an engineer to annotate it.

➡️ If a company employs a team of 20 annotators, it would still take them 600 hours to process this one specific dataset.

➡️ Now, imagine having hundreds of these datasets every month or week, where an average geospatial 3D point cloud dataset can reach from 200 million to 2 billion points.

➡️ This approach results in a linear increase in cost, where businesses inevitably have to cancel projects if they simply can’t meet the demand. The pace at which the data is created can not be matched with the capacity of current data annotation service providers.

💡That’s where machine learning can do a lot for the business.

✅ Automation is an obvious solution and we’re providing it by combining hardware, software, and state-of-the-art ML models to deliver 30x more efficient processing of the 3D point cloud.

👇🏼Find out more about Thor X, the first System-on-Cloud, and step into the future of cloud systems: https://anari.ai/thor-x/

22/08/2022

WHY DOES THOR X, THE FIRST SYSTEM-ON-CLOUD MATTER TO THE GEOSPATIAL INDUSTRY❓

💡According to Geospatial World, GW Consulting estimated the global geospatial market to be worth approximately USD 395 billion in 2021, compared to USD 365 billion in 2020, while the numbers are expected to grow at a CAGR of 14.61 percent between 2021 and 2025, to touch USD 680 billion.

👉In the future, geospatial data will hold greater significance than ever before, which requires a proper approach to data processing, especially in terms of 3D point cloud semantic segmentation.

“The quality data annotation of 3D point clouds is the key factor for numerous industries in order to provide value to the end users. AI plays a significant role in this regarding accuracy and speed, and will be crucial for semantic segmentation of the 3D point cloud in the future”, said Avi Yashar, co-founder and Chief Product Officer at Dataloop AI.

✅Thor X, the first “System-on-Cloud” approaches it with the unique combination of hardware, software and machine learning, where an AI-empowered system delivers currently the highest efficiency on the market in terms of 3D point cloud processing in the geospatial industry.

Find out more possibilities coming with Thor X: https://anari.ai/thor-x/

18/08/2022

MOST COMMON CHALLENGES IN 3D POINT CLOUD PROCESSING ⬇️

While exploring how geospatial companies process 3D point clouds, we found out that the most common obstacles come from the following challenges:

➡️ The first one is point cloud acquisition which would accurately mirror a 3D object

➡️ Visualizing the recorded system and creation of a semantic representation appears as a second problem.

Although many software approaches it, there is still a huge bottleneck when it comes to the classification of objects from the 3D point cloud datasets. This is very evident in industries that work with large point cloud datasets, such as 3D GIS, 3D CAD and BIM, which count up to tens of billions points collected in a single scan, translating to hundreds of GB or even TB of data

➡️Finally, implementing the domain knowledge in order to create a precise and reliable 3D model

💡Anari’s approach to these problems comes as a powerful combination of hardware acceleration, domain-specific algorithms, and state-of-the-art ML models. We developed Thor X to solve some of the biggest challenges the geospatial industry is facing.

✅The main output is a 30x better performance in comparison to one of the Nvdia’s best GPUs on the market, where the efficiency reflects in the throughput, price and power consumption.

Find more information here: https://anari.ai/thor-x/

17/08/2022

SYSTEM-ON-CLOUD TECHNOLOGY | The future of Cloud systems 💡

“System-on-Cloud” technology is a technology combining hardware and software architectures together with machine learning models to revolutionize the way AI compute systems are utilized.

Developed by Anari AI, the first “System-on-Cloud” - Thor X is specialized in semantic segmentation of 3D point cloud data structure by providing custom hardware acceleration to industries such as geospatial and BIM.

💡What’s the best about Thor X is that the system enables even 30x more efficient processing compared to the best GPUs currently available on the market.

👉And why does the world need it?

Because general-purpose hardware like GPU or CPU has proven effective in processing 2D images, but it falls short when met with the complexity of 3D structures.

💡On the other side, FPGA embedded Thor X is designed to specifically respond to the needs of 3D point cloud datasets obtained by both LiDAR scanning and photogrammetry.

Want to know more about why the 3D point cloud related industries need Thor X?
✅Find more information here: https://anari.ai/thor-x/

28/07/2022

The key players for 3D point cloud processing software on the global market use different approaches to tackle exacting challenges in creating a digital 3D model of a complex object in the physical world.

These challenges can be divided in three main categories:
➡️ Point cloud acquisition which would accurately mirror a 3D object
➡️ Visualizing the recorded system and creation of semantic representation
➡️ Implementing the domain knowledge in order to create precise and reliable 3D model
Regardless of many advanced software processing features, there is still a need for a huge manpower working on point cloud labeling.

👉 In order to help make these processes much more efficient, Anari AI developed a new cloud-based technology - “System-on-Cloud”, which introduces an optimized and efficient system from various hardware and software architectures combined with machine learning models.

🚀 Anari's Thor X delivers up to 30x more efficient processing compared to the best general purpose hardware available on the market, also providing many other advanced solutions that answer the mentioned problems.

✅ Read more about challenges in existing software processing of the 3D point cloud and find out how Thor X improves the process:
https://anari.ai/thor-x/