AI LabAI Lab

Welcome to the LAB!

At HACARUS we are always tinkering with new tools and techniques to stay ahead – we are an R&D driven organization, with a myriad of internal development projects running at all times. With this page we offer our partners and stakeholders a chance to peek inside these activities, to get a feel for what’s coming next, and where we see the AI field and HACARUS moving in the future.

Collaborate with us

IN THE LAB

Edge AI Platform: Lightweight and Explainable AI Solutions for Edge Devices with Complete Management from the Cloud

One of the greatest challenges facing an increasingly connected industry is the quickly growing number of deployed IoT devices. Spread across hardware makers, and software builds it is today difficult to manage and get a complete picture of all the devices and their performance. To achieve the true benefits of IoT, users must have full control of complex networks of devices, in fact data driven efficiency improvements are reliant on data harmonization and the utilization of AI solutions.

Levering our experience in edge AI development, Hacarus’ is currently building an end-to-end platform for IoT device & AI tool management – aimed at putting our customers in full control of industry 4.0 environment.

Requirements for Industry 4.0 Deployment

Seamless storage of data generated across a wide spectrum of devices

Device specific control and management of edge nodes

Data security & privacy ensured for all deployed sensors

Full control and command of deployed AI models – with the ability to re-train on the fly

Dynamic architecture and design – able to adjust to changing conditions automatically

A complete End-to-End Platform for Edge AI Device and IoT Cluster Orchestration

Developing Solutions Capable of Detecting and Treating even the Rarest Conditions

As advanced technology and AI continues to help the majority of people lead healthier lives, it’s important to consider how it can be used to diagnose and treat more complex cases like rare diseases.

In fact:
• Rare diseases affect 350 to 400 million people globally
• around 1 in 17 people will have a rare disease at some point in their lives
• In the developed world patients wait as long as 5 years for accurate diagnosis

While used as an umbrella term, as a group it is clear that this is far more common than you may think. And it’s not only patients that feel the consequences of long time to get an accurate diagnosis but also healthcare providers and caregivers.

At HACARUS our Sparse Modeling based AI technology is uniquely suitable to help doctors diagnose rare conditions, as we can work with small data sets and are able to provide explainable AI. Our current development activities are focused on building tools that allow us to provide create a comprehensive solution for early detection of rare conditions.

Understanding the Optimal Recipe - from Data Driven Insights

Increasingly data is becoming the cornerstone of material science, simulation models are widely used, as oppose observation and theories as in the past. At the same time the amount of data that has been generated by simulation models has grown beyond the analytical capabilities of conventional tools. Therefore, AI could be a key player in accelerating the pace of discovery and deployment of advanced material systems.

There are plenty of opportunities for AI and data informatics in material science as they could accelerate the research time of the development and commercialization cycle of new materials. With the help of machine learning algorithms, researchers will be able to run analysis on data from experiments and simulations to get new measurements and insights for new materials and how-to synthesize them.

One important function of AI in the domain of material science is prediction of material properties. Tests application and experimentation can generate a huge amount of data. Using machine learning it is possible to extract underlying trends and patterns within these data sets and predictively map features (i.e., materials descriptors) to target materials properties.

At HACARUS we are currently exploring tools and methods to apply our lightweight, explainable AI solutions to this compelling field.

    

WHITE PAPERS & PRODUCT MATERIALS

Click an item below to view and download

Sparse Coding Implementation for Intel FPGA Devices

Download

Using Sparse Modeling in AI: A Human Centric, Explainable Approach

Download

Explainable AI

Download

What Challenges Does Sparse Modeling Solve?

Download

Less is More

Download

Environmental Impacts of AI

Download

Capturing the first ever image of a Black Hole – How Sparse Modeling made it possible

Download

TECHNICAL BLOG

AI MODULES

Our AI technology is provided in the form of a single function called a module. Modules are suitable for bringing in only certain AI functions into an existing system. By combining modules, you can also cope with more sophisticated and complex issues. The module can be used via cloud API or SDK for embedded devices. We offer a wide range of modules to meet various industrial and medical applications.

Motion Detection

Motion Detection

Detection of moving objects, building blocks for many video analytics scenarios.

Prediction

Prediction

Analyze past data and predict future trends and events.

Anomaly Detection

Analysis of sensor or biological data with alerts when anomalies are detected.

Scoring

Analyze, categorize and rank user attributes and historical data.

Super Resolution

Super Resolution

Generate high resolution images from low resolution images.

Face Recognition

Face Recognition

Recognizes human faces from images and videos. Use cases include visitor statistics and surveillance.

Classification

Classification

Analyze a wide variety of data and categorize it into groups and categories.

Recommendation

Recommendation

Analyze user attributes and historical data and generate recommendations.

Inspection

Inspection

Identify defects from images or video streams.

Vascular Constriction

Vascular Constriction

Identify where the blood vessels are constricting in the MRI / MRA / CT image.

[Cloud]

Treatment Suggestion

Treatment Suggestion

Proposing treatment plans according to pre-specified guidance and criteria.

[Cloud]

Disease Classification

Disease Classification

Analyze patient profiles, diagnosis results, and biological data, and classify disease types.

[Cloud]

ECG

ECG

Analyzes ECG (electrocardiogram data) to detect abnormalities and predict the occurrence of diseases.

[Cloud] [Embedded] [Chip IP]

Subscribe to our newsletter

Click here to sign up