Specialized microchips to reduce latency in objects using AI
By Aarish Technologies | Published on March 11, 2022
Problem and objectives
Problem
- Use of AI requiring significant computational effort;
- A cloud-based system is required to operate technological tools such as facial recognition cameras or self-driving cars.
Objectives
- Reduce the quantity of electricity used by microchips.
- Enable AI techniques to be implemented in a variety of technological objects;
- Increased system performance;
- Reduction in data security risks.
Proposed solution
Solution
Aarish Technologies creates processors for computers with a high processing speed. The product’s long running time and performance make it possible to help implement artificial intelligence (AI) techniques in all types of technological objects. The tool helps avoid system latency and speeds up the computing capacity of computers in which it is installed. It also allows an AI system to be directly installed in a sensor without going through a cloud system.
Principle
Aarish Technologies creates microchips that fundamentally alter the capacity to integrate AI tools in technological objects. The solution patented by the company reduces the computational load of convolutional neural networks by 70-90% and avoids having to rely on a cloud-based network to keep detection tools running properly.
Role of AI
The technology developed by the company facilitates the use of AI in a variety of applications. Some fields in which it works include:
- Algorithms;
- Segmentation;
- Classification models;
- Image analysis;
- Image classification;
- Image recognition;
- Software.
Impacts
Main outcomes
- Reduced costs for businesses that use AI;
- Reduction in the quantity of energy used by AI systems;
- Ability to run algorithms directly through the detection tool (e.g. camera) without using a cloud system.
Generation and dissemination of new knowledge
Patent held by the company.
Conditions for success
Team mobilised
Aarish Technologies’ AI team consists of:
- Data analysts;
- Solution architects;
- Data scientists;
- Data engineers;
- Software developers.
Project stages
Design and manufacture processors.
Project stages - to come
Market the product;
Create partnerships.
Context of the solution
Use cases
Examples of applications in the automotive industry include:
- Object detection systems able to operate in low visibility and rural environments;
- Greater precision in object detection and decision making.
Examples of camera facial recognition applications include:
- Direct detection by the camera without having to use a cloud system;
- Reduction in costs;
- Increased efficiency and reduced system latency.
Customers
Aarish Technologies’ clients include:
- General public;
- Consumers;
- Workers;
- Companies;
- Governments;
- Educational institutions.
The company’s technology can be used in a multitude of industries such as:
- Mining, quarrying and oil and gas extraction;
- Utilities;
- Construction;
- Manufacturing;
- Retail trade;
- Transportation and storage;
- Professional, scientific and technical services.