Human learning with machine learning: the game of Go and real-life applications
When working with machine learning tools and algorithms, we often assume that only the machine is learning. Not so, argues Mark Hsieh. To achieve maximum value, we need to learn from the results generated by algorithms to better understand the problem we are trying to solve, for example in predicting disease progression in diabetes.
Satellite IoT for 5G — What’s the Story?
What does the rise of 5G and new (LEO) space operators mean for the satellite and cellular industries and for next-generation IoT? TTP’s Adrian Hillier charts the way ahead.
Evolution in user terminal design for broadband from space
New LEO satellite networks promise high-speed broadband access for billions living beyond the reach of fibre and cellular networks. But the cost of User Terminals (UTs) for such satellite connections remains far beyond consumer budgets. Andrew Fell sketches how technical innovation, convergence in standards and network intelligence could pave the road to low-cost UTs for broadband from space.
Smart antennas for the IoT: for the few, not the many?
Smart antennas can radically improve IoT network performance - and they can do this with no changes to the population of wireless IoT devices. Only the IoT gateway terminals – usually relatively few in number – need the upgrade. But ultimately everyone reaps benefits – network operators from improved coverage, data capacity, network resilience and location services, and network users from lower cost and lower power IoT devices.
Could technology help contain the next equine flu epidemic?
Humanity has endured infectious disease epidemics for as long as it’s existed. However, as global connectivity grows so does the rate at which they can spread and the number of people they can affect. The origin of epidemics is the random genetic mutation of pathogens so we have no control over their arrival but technology offers very real control over the way they spread.
How to build working AI models when you are short on training data: AI as part of the business transformation toolkit (Part 3)
The sights of Artificial Intelligence (AI) can be trained on almost any business problem that will benefit from automation. But what if training data for your problem is in short supply? In this blog, Simon Shakespeare shares a few tricks for training AI without reinventing the algorithmic wheel or by creating training data out of thin air.
New modes of drug action tackle intractable diseases
2018 was a busy one for the US Food and Drug Administration. Drug approvals hit a record high, with 59 novel drugs securing the agency’s permission to head to market. This included a number of drugs with modes of action that transcend the two main classes of drugs.
No pain, no gain: R&D challenges for the success of microneedle systems
Microneedle systems pose multidisciplinary R&D challenges—some because of and some despite the much smaller size of the needles involved. But microneedle systems also offer exciting new prospects for drug delivery and multiplex biomarker detection. This information could be used to guide personalised treatment or enable value-based healthcare.
The future of molecular diagnostics: Will isothermal amplification techniques steal PCR’s crown in the close-to-patient era?
PCR is still the go-to method in the molecular diagnostics industry but, with the rise of close-to-patient testing, isothermal amplification methods with their low power and potentially easier detection schemes look increasingly attractive.
4 Benefits Satellite Operators and their Customers gain from an Edge-Centric Approach to the Industrial IoT
In our concept paper "Satellite at the Edge", we extolled the virtues of an edge-centric approach to Satellite IoT. Having described the “what”, we now set out the “why” by outlining four major benefits and asking whether they primarily advantage satellite network operators or their customers.