Jeffrey J. Bloom

Jeffrey J. Bloom

While caffeine is easily the most abused substance, young entrepreneurs are now "body hacking," aka using performance enhancing meds--like modafinil (wakefulness), methylphenidate (concentration) & nitrazep-am (anxiety & insomnia relief)--that promise cognitive benefits & improved concentration for executing complex tasks or putting in longer hours.

However, users dread the ensuing crash when the drugs wear off. Entire online forums are devoted to procuring drugs & discussing side effects.

“Stopping usage leads to withdrawal symptoms ... the person will experience a crash ... start to feel dull & have zero motivation,” said Dr. Benegal.
http://www.thehindu.com/sci-tech/health/body-hacking-for-an-edge/article19820540.ece

Jeffrey J. Bloom

Autodrop3d is an open source system that solves the problem of needing a human to remove a 3D print from its print bed. Implemented as an open source hardware and software system, it allows for web based, multi-user print queue, automatic notifications, and web-based CAD design tools to all be integrated in one open source system. There's a video that shows the hardware in operation and a link to the web site with a Git repository for the software and hardware components.

Autodrop3D is now raising money on Kickstarter, promising to show their support for open source innovation by "releasing all of our documentation, design files, and software prior to the end of this Kickstarter campaign."

Jeffrey J. Bloom

A PABX/PBX (Private (Automatic) Branch exchange) is telephone equipment that is installed on corporate premises to provide a number of telephone extensions within an offi¬ce and operate as a connection between the business and the external dial-out network. A PBX allows sharing of outside lines and thus significantly reduces the number of such lines needed to be leased from a Telco. PBXs have evolved over time in much the same way that Telcos have evolved, moving from circuit switched to IP-based packet technologies. This Point of view on PBX hacking takes a look at some of the most common methods of hacking PBXs and recommends the preventive steps to counter them.

Jeffrey J. Bloom

Scientists have unravelled the complexity of in-built sat-navs in bees' brains. Consisting of a complex network of neurons integrating every detail of outbound journeys, they create memories used to plot direct routes back to their hive.

A detailed computer model of a bee's brain was created by studying nerve function via electrodes & microscope images, which could lead to simple lightweight autonomous navigation algorithms not requiring GPS.
https://phys.org/news/2017-10-brain-reveals-insects-beeline-home.html

Jeffrey J. Bloom

"Screwdriving," exploits the bluetooth low energy (BLE) protocol, often used by “smart” sex toys, as demonstrated by security consultant Alex Lomas. He also hacked his father’s BLE-enabled hearing aid, altering hearing & putting voices in his head. In February, a teen hacked 150k printers to raise awareness of device exposure online. The FDA recently recalled almost half a million pacemakers due to fears they were vulnerable to hacking.

The internet connects people & IoT connects everything else; your fridge, car, home, cuddly toy, bus, nuclear power station & yes, sex toys. It's happening now ninjas.
https://www.theguardian.com/commentisfree/2017/oct/06/drive-by-sex-toy-hacking-wake-up-call-britain-internet-security-vibrators

Jeffrey J. Bloom

If you're still unsure about whether you should take GodMode gamer pills, definitely talk to your doctor. If that conversation goes well and you still aren't sold, Boss Level Labs provides a list of "pro users" of its product, which includes a handful of game developers, people who lift weights, someone who is a "celebrity financial advisor and motivational speaker," and a retired Hungarian chess grandmaster named Judit Polgár.

Outside of the benefits you'll undoubtedly feel while gaming, I guess, the website says GodMode "should help you remember where you put your keys."

Jeffrey J. Bloom

Relying on algorithms without understanding their logic is like script kiddies never learning to code. The math is often unpure, according to mathematician Cathy O’Neil.

Analytic algorithms may be wrong & are reinforcing biases, erroneous information & even their developer's bias. "Algorithms are opinions embedded in code ... we’re hiding behind mathematics as a shield."

"People who create algorithms embed their own definitions of success ... Machine learning doesn’t make things fair, it represents past patterns & automates [them]."
https://www.rtinsights.com/analytic-algorithms-may-deliver-bad-data-at-real-time-speeds/

Jeffrey J. Bloom

“It’s been interesting to note in the last couple of years, the trajectory of AI and what it’s done for us that kind of justifies the investment,” said Paikeday. “Like, deep neural networks have been growing; neural network development has been more and more computationally expensive as these networks get smarter, and by consequence bigger … [it’s the] GPU that’s allowed these models to get so big, so extensive, so computationally expensive, because we’re now providing the compute capacity to execute these massively parallelized algorithms that power the biggest neural networks that there are today, with speed that really wasn’t possible until developers way back when started training their models on GPUs instead of CPU servers.”

Jeffrey J. Bloom

Rasa Core is available NOW in open source via GitHub! Instead of using Amazon's Mechanical Turk to correct ML models, they built a platform allowing anyone to train & update models via sample conversations.

Context is everything with dialog systems. We take for granted the complexity of even simple conversations. Rasa’s approach allows customers to bootstrap models without training data, only a few dozen sample conversations are needed to get a bot working effectively.
https://techcrunch.com/2017/10/04/rasa-core-kicks-up-the-context-for-chatbots/

Jeffrey J. Bloom

From entire AI platforms that handle every possible task you may need, to one-off apps that recognise speech using ML, AI is all around us & should be celebrated.

1. Microsoft’s Azure Machine Learning Studio is a one-stop-shop for developing your own AI apps & services. Designed for ease-of-use, pre-determined algorithms help you implement AI into whatever you’re creating. There's also a load of machine learning resources available to get you up to speed in no time.

2. Google’s Cloud Speech API is available for developers with a remarkably low 4.9% word error rate. With recognition of 110+ languages & variants it converts speech to text more accurately than any other speech platform.

3. Inbenta is innovating intelligent chatbot technology, using natural language processing & AI to find out the meaning behind search queries & allow for better responses. It’s simple to implement & provides in-depth reports so you can see results.

4. Anodot analyses large datasets & finds anomalies using AI, turning them into key business insights. It makes use of its own patented ML algorithms, isolating potential issues so humans can examine them & take action if needed.

5. H2O.ai makes it possible for anyone to create a model of their business/industry, using prexisting data & ML. You can start generating valuable insights right away on a large scale. Easily slotted into Python, Scala & R environments too, making it an extremely flexible platform for predictive data analysis.
http://www.itpro.co.uk/cloud/29590/best-ai-software-and-technologies

Jeffrey J. Bloom

Although ITSM solutions are rapidly evolving, service management will never go away as long as IT exists. However, by implementing AI technology, IT service management will experience a disruptive change that will alter the way humans are involved with the service management process. It is also important to note that these disruptive changes will affect all service management operations, not just IT service management.

Jeffrey J. Bloom

The way I think, is how I code. I need to converse with those who understand the purpose & goals, before reviewing data sets, notes, binders, xls, doc, pdf, etc.. That's how I shape my frame of reference & perspective, because Machine Learning (ML) algorithms need context to "learn."

Data wrangling currently accounts for ~80% of time spent developing data analytics, due too messy data sets. Automating the process is key, combining data wrangling & analysis into ML algorithms has great potential. But it's a chicken or the egg situation: Is it best for ML to learn from raw data, or train ML to learn desired outcomes before learning from data?

Lives & economies will be transformed by data science & AI. There are challenges in data wrangling as a starting point & in AI as an end point. The future is here, as seen in many retail & healthcare implementations.

When properly implemented, systematic data use benefits every process & as the degree of integration increases, so do possibilities. Of course none of this happens without effective data infrastructure.
https://www.prospectmagazine.co.uk/politics/the-future-will-be-data-driven

Jeffrey J. Bloom

In 2016 Google's AI had an IQ of 47.28. It came out ahead of Chinese search engine Baidu (32.92), Microsoft's Bing (31.98) & had almost double the IQ of Siri (23.94).

Notably, none of these systems had a higher IQ than a 6-year-old (55.5), much less an 18-year-old (97), the researchers found.
https://www.cnbc.com/2017/10/02/google-ai-has-almost-twice-the-iq-of-siri-says-study.html

Jeffrey J. Bloom

AI, machine learning (ML) & even AR & VR, can help manage the onslaught of data from multiple sources into the big data ecosystem to properly gather, mine, prep, store, analyze & help smooth operations.

"The data lake isn't where data goes to die," Gartner analyst Merv Adrian says, "it's where data goes to live." As simplistic as it sounds, it's up to data managers to move beyond traditional data control & into the critical task of improving data quality & delivery -- taking all that raw data & making it useful.

Semantic technology could be a blessing or curse to AI; companies are gearing up for greater big data management deployments & all data must be treated equally in the search to find value.
http://searchdatamanagement.techtarget.com/ezine/Business-Information/Managing-the-big-data-ecosystem-requires-agility-amid-disruptions

Jeffrey J. Bloom

"The field of metagenomics needs standards to ensure that results are correct, well validated & follow best practices," Pop said. "For instance, if a doctor is going to stage an intervention based on results from metagenomic software, it's essential that those results be correct. Our work provides a roadmap for choosing appropriate software."

Jeffrey J. Bloom

The timing of the new connection is particularly curious as US CyberCom was carrying out DoS attacks against state-sponsored hackers in North Korea over the weekend. The attacks were designed to overwhelm the hackers connectivity make their systems impossible to use.

The Russian internet connection to North Korea began less than 24 hours after CyberCom attacks were scheduled to end & is likely to provide the nation with additional protection in case of future DoS attacks.

Jeffrey J. Bloom

Equifax says about 2.5 million additional consumers may have been impacted by last months cyber attack. Their initial disclosure said personal details of up to 143 million consumers were accessed by hackers between mid-May and July.

As for what led to the breach, Ars Technica reports "a series of costly delays & crucial errors." Chief among them: an Equifax e-mail directing admins to patch a critical vulnerability in the open source Apache Struts Web application framework went unheeded, despite a two-day deadline to comply. Equifax also waited a week to scan its network for apps that remained vulnerable. Even then, the delayed scan failed to detect that the code-execution flaw still resided in a section of the sprawling Equifax site that allows consumers to dispute information they believe is incorrect. Equifax said last month that the still-unidentified attackers gained an initial hold in the network by exploiting the critical Apache Struts vulnerability
http://www.reuters.com/article/us-equifax-breach/equifax-failed-to-patch-security-vulnerability-in-march-former-ceo-idUSKCN1C71VY

Jeffrey J. Bloom

The Holographic Brain Project is an application run on Microsoft's augmented reality Hololens headset. It projects an interactive 3D brain into the user’s physical environment with brain structures that can be enlarged, isolated or rotated, among many other commands.

Jeffrey J. Bloom

Not to worry, soon AI will accurately tell us the probability of the actual sender writing a message...

Graeme had not been copied in & the smiley faces, normally peppering Giles' emails, were suspiciously absent.

She guessed the scammers identified a large payment was due by reading emails between Giles & Marlborough Vintners over the month prior.

Bank staff identified the account was subjected to "genuine infiltration" by offshore hackers.

Jeffrey J. Bloom

NVIDIA is partnering with Alibaba--adding to 50+ major tech companies using Metropolis--helping them tap into one of the largest smart city markets on the planet: China.

Launching in May, NVIDIA's Metropolis cloud platform sifts through surveillance & traffic video using AI to manage traffic, parking, law enforcement & other city services.

Metropolis uses several NVIDIA technologies -- including Tesla GPU accelerators, deep learning software & DGX-1 cloud-based supercomputers -- to comb through the vast amount of data collected by a city's security & traffic cameras to make decisions based on what it sees. The platform can monitor video in real-time & view data in video recordings up to 30 times faster than humans can in some cases.

NVIDIA believes that 1 billion cameras will be in service in cities around the world by 2020 & Metropolis' AI will allow local governments to reduce traffic & crime while increasing efficiencies block by block.
http://host.madison.com/business/investment/markets-and-stocks/nvidia-has-its-eye-on-smart-cities-in-china-with/article_93d8da39-6e31-5fb1-8009-d9ad41933f5b.html

Jeffrey J. Bloom

Insiders say many of the tools used in the hack were Chinese in origin & there are similarities to China-backed breaches targeting the health insurance firm Anthem & the US government's OPM. Also, none of the stolen data has surfaced online -- whoever took it wasn't in a rush to profit. This was a "'get as much data as you can on every American' play," one of Bloomberg's contacts said.

However, sources aware of federal investigations say there's evidence of a state-sponsored attack, not that it points to any one country. Equifax's security consulting partner, Mandiant, wrote as recently as September 19th that it couldn't identify the culprits or where they came from.
https://www.engadget.com/2017/09/30/equifax-breach-shows-signs-of-possible-state-sponsored-hack/

Jeffrey J. Bloom

Existing quantum computers can only handle as many as 12 qubits, the maximum ever manipulated was by Russian researchers which amounted to 51 qubits. Japanese researchers believe their circuit is capable of processing, a million qubits. However, they're still translating their theory into reality. Furusawa said "We’ll start work to develop the hardware, now that we’ve resolved all problems except how to make a scheme that automatically corrects a calculation error." It goes without saying that this possible breakthrough could help solve many great problems of today which include medical research, transportation, electronics & the economy in general.
http://wallstreetpit.com/114245-scientists-invented-quantum-computing-technique-process-millions-qubits/

Jeffrey J. Bloom

Kodi's popularity & its plugins providing streaming media, have made it the latest target of Hollywood's copyright enforcement juggernaut. The difficulty facing the TV titans is that neither Kodi sellers, nor those who write or host add-ons, are engaging in unauthorized copying, cases targeting these parties have to rely on other legal theories. These lawsuits seem to have a few goals: to expand the scope of secondary copyright infringement yet again, to force Kodi add-on distributors off the Internet & to destroy open source, freely configurable, media players by focusing on a few bad actors.
The EFF details the specific lawsuits in each region & concludes that their courts "should reject these expansions of copyright liability & TV networks should not target neutral platforms & technologies for abusive lawsuits."

Jeffrey J. Bloom

A 300% increase in AI investment this year? Time to get smart! Here's your top 10 AI & ML (Machine Learning) tools, direct links located at bottom of post:
1. Microsoft Azure - built on ML capabilities of several MS products & services, like Cortana, it also uses proven solutions from Xbox & Bing.
2. AI-one’s "Analyst Toolbox" provides a library, building agents & APIs for developers & can turn data into rule sets, enabling in-depth AI & ML structures.
3. DiffBlue builds an exact mathematical model of any code base, allowing deep semantic understanding of a program's purpose. Its aim is automagically finding & fix weak code.
4. Google’s TensorFlow uses data flow graphs with nodes = math operations & graph edges = multidimensional data arrays (tensors). The flexible architecture deploys computation to one or more device with a single API.
5. Amazon Web Services has 3 AI toolkits: Rekognition adds image interpretation & facial recognition, Polly automates voice to text across 47 voices in 24 languages, & Lex is the Alexa engine, allowing chatbot integration.
6. Protege's suite is ideal for "knowledge-based apps with ontologies" & lets developers create, upload, modify & share apps in an active community, making troubleshooting & collaboration easy.
7. Apache Spark MLlib is a memory cluster based data processing system designed for simplicity, scalability & easy integration with other tools.
8. Nervana Neon - Python-based ML library lets developers build, train & deploy deep learning tech, with lots tutorials & a 'model zoo' housing pre-trained algorithms & example scripts.
9. OpenNN - formerly known as Flood it's aimed at experience developers. It's a source class library in C++ for neural networks, with lots of documentation & tutorials.
10. Apache Mahout - a library of scalable ML algorithms, built on Apache Hadoop & MapReduce paradigm. The project aims to make it faster & easier to turn big data into big information.

1. Microsoft Azure ML: https://azure.microsoft.com/en-in/services/machine-learning-studio/
2. ai-one: http://www.ai-one.com/
3. DiffBlue: http://www.diffblue.com/
4. Google’s TensorFlow: https://www.tensorflow.org/
5. Amazon Web Services: https://aws.amazon.com/machine-learning/
6. Protege: https://protege.stanford.edu/
7. Apache Spark MLlib: https://spark.apache.org/mllib/
8. Nervana Neon: https://www.intelnervana.com/neon/
9. OpenNN: http://www.opennn.net/
10. Apache Mahout: http://mahout.apache.org/

Jeffrey J. Bloom

Now a single operator using emerging human-brain interfaces can control a swarm of drones, making possible new classes of applications, according to Panos Artemiadis, director of the Human-Oriented Robotics and Control (HORC) Lab at Arizona State University.

Jeffrey J. Bloom

The Neural Engineering System Design (NESD) program aims to develop systems that can electronically communicate with the region of the brain that assists sensory function. In July, DARPA awarded $21.8 million to 6 teams: 4 will focus on vision & 2 will focus on hearing & speech. By next summer, only a few teams will move on to phase 2. By the end of the four-year program, DARPA hopes the tech will mature enough to read activity from 1 million neurons, stimulate 100k neurons & provide closed-loop communication with the brain.

"If you set the bar high, sometimes you get across it," NESD director Brad Ringeisen said. "That’s been the special sauce about DARPA throughout its history." The goals may be lofty for the time frame, but hopefully the tech will advance enough for other organizations to continue what we start here.
https://www.bizjournals.com/washington/news/2017/09/29/darpa-exploring-new-possibilities-in-using-tech-to.html

Jeffrey J. Bloom

"In the future, you've got transportation, health care, entertainment. All of these will change in amazing ways & create trillions of dollars of economic value. So you also have this war for AI architecture leadership that probably will be the biggest battle of our lifetime," Applied Materials CEO Gary Dickerson said.

Volatility has decreased across industry in the last seven years & that the "war" for leadership in mobile has increased the need for top-performing chips. "The materials that create the power & performance [for chips] come from Applied Materials." So in that war, Applied wins no matter what.
https://www.cnbc.com/2017/09/28/applied-materials-ceo-on-the-war-for-ai-leadership-in-semiconductors.html

Jeffrey J. Bloom

Waves Glasses are a brand new pair of glasses that are both ocular and device-friendly. As glasses, they lower the effect of screen lights on your eyes and can transition to protect them from UV rays. They can also fit any prescription lenses so that anyone can use them. But as a tech tool, they can do so much more.

Waves can connect to any of your bluetooth-enabled devices for a hands-free experience when you're on the go. By tapping on the side of your glasses (think Cyclops from X-Men without the laser eyes), you can answer calls, skip through your playlist, or even ask Siri to cover something for you, all without ever touching your devices.

Jeffrey J. Bloom

Machine learning consists of many simple logic operations needing connectivity, not slow complexity of current CPUs. "Convolutional neural networks" involve recursion, feedback & simultaneous computations feeding into each other. "You want something that is simple on compute, on arithmetic & very intense on comms," says Feldman. A new chip is needed to handle sparse matrix math & emphasize comms for I/O of calculations.

"Workloads are changing dramatically," due to machine learning, "whenever workloads have changed in computing, it always creates an opportunity for new kinds of computing."

"There's going to be a vast new wave of innovation in computer chips," says Feldman. Machine learning is changing the requirements for computer chips.

Optimized for complex logic operations, "CPUs are remarkably poor" at communicating results between CPUs.
http://www.barrons.com/articles/cerebras-founder-feldman-contemplates-the-a-i-chip-age-1506629165

Jeffrey J. Bloom

Increased product durability & lifetime? Yes please! Imagine not worrying about corrosion or environmental wear & tear on buildings & vehicles.. We're already seeing it, with increased applications of practical Self-Healing Materials across most industries.

The acceptance & proliferation of self-healing material, not to mention its untapped potential, are enough for me to keep my eye on this growth Market.
https://www.openpr.com/news/739514/Self-Healing-Material-Market-is-expected-to-grow-over-the-CAGR-of-around-7-during-the-period-2017-to-2023.html