New Zealand based AI startup relieving people of repetitive, unsatisfying work

A New Zealand business that makes its own servers is moving into AI, providing machine learning solutions that do routine tasks and relieve workers of monotonous, unhappy employment. Kauri cone has lately shifted from developing low-cost, high-efficiency, ecologically friendly hardware to inventing software that solves commonplace issues for us. These jobs include recognizing rubbish for enhanced recycling, 'looking' at road things for automated safety, pest detection, and counting sheep.

Mike Milne, Kauri cone's managing director and founder, claims the company makes application, cluster, and IoT servers. In this area, machine learning at the network's edge emerges.

According to Mike Milne: Having built low-cost-low-power edge technology, we saw a significant possibility for smart computing in certain not-so-enjoyable daily tasks. It is good for gear, programming skills, and decent mobile network coverage.

There might be as many use cases as tedious but important activities. AI can monitor hundreds of kilometers of road and it costs. Construction waste is widespread, harmful, and can be effectively handled.

Work is doing things we'd rather not or can't do ourselves 

Work is doing things we'd rather not or can't do ourselves. Despite living in a fantastically evolved era, there are daily jobs that don't demand much attention or knowledge.

Image recognition is more complicated than it seems. Because recognizing items is challenging for machines, 'Captcha' problems commonly include identifying traffic lights. ML solves the difficulty by 'training' on thousands of photos to learn hits and misses.

IoT server hardware was used to solve each difficulty

Kauri cone's RISC IoT server hardware was used to solve each difficulty. Solar-powered edge devices require as little as 3 watts and operate forever on a single solar panel, running Ubuntu and Python or other open-source languages. Milne believes this creates a dependable, low-cost 'field-ready' gadget.

The Rocks on Roads study showed the problems of simple picture recognition. Kauri cone ran a training model around the clock for 8 days, accumulating 35,000 rock photos, which extended to 3,000,000 recognizable features (bear in mind, a human identifies a rock almost instantly, perhaps faster if hurled). This training helped the system locate road rocks.

The Kauri cone IoT server will monitor construction waste kinds and volumes for a new project. Trained to recognize trash kinds, the generated data will improve garbage management, recycling, and responsible disposal.

Counting sheep is a vital duty for New Zealand farmers. As an ML exercise, it predicts improved stock management, as does Kauri cone's pest detection test case. The ever-watchful camera and related electronics detect, number, and monitor grass levels for sheep. This small-scale application is ready for expansion.

Results

Milne thinks Kauri cone's modest test cases are only the beginning and predicts widespread ML applications. According to him, there's no limit to the amount of everyday jobs where computer vision and ML can reduce our effort and enhance efficiency.

Rocks on Roads offers enhanced safety with reduced 'human' overhead, decreasing or eliminating human mistakes. Waste management is a multifaceted problem where the employment of personnel is difficult due to simple economics (and potentially stultifying work); New Zealand's primary sector is ripe for technologically powered performance improvements that could boost productivity through automation and improved control; and pest management can help the Department of Conservation and allied parties achieve better results using fewer resources.

With developing cellular and low-power networks like SIGFOX and Lora Wan, infrastructure is accessible even in rural areas. Low-power hardware takes computation to the edge. Now it's a question of finding opportunities and submitting applications