Bringing artificial intelligence to mobile computing is a significant challenge. That’s the goal of Qualcomm’s new Zeroth Processors.
Mimicking the human nervous system and brain to allow computers to learn about their environment and modify their behavior based on this information has long been the goal of artificial neural networks. Whatever computing model is used to achieve this capability the real problem is one of scale. The human brain is estimated to have 100 billion neurons — with 100 trillion connections. That is at least 1,000 times the number of stars in our galaxy.
These computational models can be implemented in software (e.g. Grok), but the ability to scale to the levels required for even simple human-like interactions is severely limited by conventional computing platforms. The Zeroth Neural Processing Unit (NPU) is a hardware implementation of the brain’s spiking neural networks (SNN) method of information transmission. Integrating the NPU into computing platforms at the chip level would begin to address the computational and power requirements for these types of applications.
The goals of the Zeroth* platform are:
Achieving ”human-like interaction and behavior” is an ambitious goal, but it seems like this is a good first step.
UPDATE (25-Oct-13): Good overview here: Chips ‘Inspired’ By The Brain Could Be Computing’s Next Big Thing.
* The name Zeroth comes from the science fiction Three Laws of Robotics. The First law was that “A robot may not harm a human being.”
Asimov once added a “Zeroth Law”—so named to continue the pattern where lower-numbered laws supersede the higher-numbered laws—stating that a robot must not harm humanity.
We’ll have to wait and see, but let’s hope so!
The FDA has issued their final guidance on mobile medical applications: Keeping Up with Progress in Mobile Medical Apps. The guidance document (PDF) will “give mobile app creators a clear and predictable roadmap to help them determine whether or not their products will be the focus of FDA’s oversight. ”
The regulatory approach is as you would expect (my highlight):
FDA intends to apply its regulatory oversight to only those mobile apps that are medical devices and whose functionality could pose a risk to a patient’s safety if the mobile app were to not function as intended.
There are six categories of mobile applications listed that the FDA intends to exercise enforcement discretion:
If a mobile application is considered a medical device it will be classified as such – class I (general controls), class II (special controls in addition to general controls), or class III (premarket approval) — and the manufacturer will be required to follow Quality System regulations (which includes good manufacturing practices, §820.30) in the design and development of that application.
For any organization that is not already under FDA regulatory control, this is a big deal. Given that there are 1000′s of medical applications already out there, even this limited scope approach will likely affect many companies. More information is here: Mobile Medical Applications.
The guidance includes many examples (including mobile apps that are not medical devices) and an FAQ.
This is a follow-up to Interoperability: Arrested Progress.
The FDA has recognized voluntary interoperability standards for medical devices: Improving Patient Care by Making Sure Devices Work Well Together.
The FDA and HHS has (my highlight):
published a list of recognized standards for interoperability intended to assist manufacturers who elect to declare conformity with consensus standards to meet certain requirements for medical devices.
The standards are searchable here: Recognized Consensus Standards. The Software/Informatics area currently lists 54 items.
Some consider this a “landmark announcement” (see here) but “voluntary” and “elect to declare” just seem like more of #1 (same old, same old) to me.
UPDATE (9/4/2013): More here: FDA Updates List of Recognized Standards, Confusion Ensues
When it comes to the state of interoperability in the medical device industry there couldn’t be a better metaphor than Arrested Development*. A dysfunctional family made up of eccentric well-meaning personalities each doing their own thing, oblivious to each other and the rest of the world.
Healthcare Un-Interoperability has long been one of my favorite subjects. Let’s review where things are these days.
The article Health IT Interoperability: How Far Along Are We? provides a nice summary of the current state of HIT interoperability. This is particularly important because:
hospitals using basic EHR systems tripled from 12.2 percent in 2009 to 44.4 percent in 2012
For better or worse, it’s the monetary incentives of the Affordable Care Act that push doctors to electronic medical records and is the primary reason for the accelerated rate of EHR adoption. The goal of having more electronic health records is to improve the quality of patient care. Reduction of medication-related errors is a great example: Lack of Interoperability has Ownership for Medication Errors. The rapid uptake of these systems can also present problems. For example, in Emergency Departments: EHR systems pose serious concerns, report says.
Nevertheless, it’s clear that electronic medical records is the future in healthcare data management. The down-side of this growth from an interoperability point of view is that there are that many more systems out there that don’t talk to each other!
Initiatives like CommonWell and Healtheway are moving in the right direction but are just getting off the ground. Also, these types of efforts are often far removed from the medical device industry and have little impact on software development and interface decision making.
Let’s step back and look at the HIMSS definition of Interoperability:
For all practical purposes only #3 (semantic) has value when in comes to the exchange of data with a medical device. As noted in Interoperability: Not a non-issue:
Semantic interoperability continues to be a major challenge and, if not addressed, will have a serious impact on the quality of care.
The same point is made here: Interoperability vs Health Information Exchange: Setting the Record Straight. Just because you can send it (exchange) doesn’t mean the recipient can understand it (interoperability).
One area that is always part of this discussion are standards. It’s unfortunate that due to technical and (mostly) non-technical reasons the following is often true:
… success depends upon a single large vendor assuming leadership. Interoperability will follow, but it won’t be designed by committee.
The distress of a well-focused cloud computing API involving a hand-full of vendors makes the outlook for HIT interoperability look particularly bleak. To make matters worse, the use of OSS in FDA regulated products face additional challenges that are not even seen in most other industries (see Open Source Medical Device Connectivity).
This is all good news for businesses that provide products and services that fill the connectivity gap. Companies like Capsule, iSirona, and Nuvon are many times the only effective way to provide an integration solution to a large number of customers.
I should note that there are some bright spots in the interoperability landscape. For example, the Continua Health Alliance has successfully pulled together over 200 companies to create a vision for inter-operable personal connected health solutions. Also, the West Health Institute is building standardized communication protocols into their embedded software for medical devices. These and numerous other successes provide hope, but are still just the tip of a very large iceberg.
Dr. Julian Goldman sums up the current situation in Medical Device Interoperability: A ‘Wicked Problem’ of Our Time:
Our years of work on medical device interoperability have led us to see the barriers (including technical, business, liability, standards, and regulatory factors) as “wicked problems,” in which there is little agreement about “the problem,” no agreement on a solution, and problem solving is complex because of external constraints.
Others (Is HIT interoperability in the nature of healthcare?) see the proprietary business model of major HIT companies as the primary culprit.
So what are some possible scenarios for the future?
At least for the foreseeable future it looks like #1 (insanity) is going to prevail. If I’m missing some huge game-changer, please let me know.
In the mean time, let another episode begin!
*No deep meaning here. Certainly not like Arrested Criticism. I’m just comparing the medical device industry to a bunch of fictional crazy people.