Are A Conscious Artificial Intelligence & Smart Robots Possible?
It would be like teaching a kid by showing a picture of a horse and then a
rhino, and then telling him a unicorn is something between these two, so he
could mostly identify it without having seen an actual picture before. So the
machine would be programmed such that it does not erase the earlier data also
known as “catastrophic forgetting” but like the brain have the capability of
”continual learning” by selective activation of cells & overlap networks,
and rather use the information to analyse the next dataset or “transfer
learning”. Moreover, efforts are underway to teach the machine by just one or
two examples, and not the millions of correct examples needed earlier which made
the data computation very humungous and actually limited the capability of the
machine. Human beings can multi-task effortlessly – can switch efficiently
between frying an egg, working in an office, playing badminton and writing
music, without compromising each of these activities individually. The UChicago
researchers have developed “context-dependent gating” and “synaptic
stabilization”, entailing activation of random-only 20 percent of a neural
network for each new task, a single node may be involved in dozens of
operations; thereby learning as many as 500 tasks with only a small decrease in
accuracy.
New phishing campaign targets taxpayer credentials
The scam could result in steep financial losses for taxpayers. Last year alone,
the IRS identified more than $2.3 billion in tax fraud schemes. The new
infection process is designed to evade antivirus tools and tricks targets into
installing the malware via a tax-themed Word Document containing a malicious
macro that downloads an OpenVPN client on the targeted machine. The malware
dropper establishes a connection to the legitimate cloud service “imgur” and
downloads the NetWire or Remcos payloads by way of a technique called
steganography, where the malicious code is hidden within an innocuous looking
jpeg image file. ... The malware includes a variety of functions including the
remote execution of shell commands on the infected machine, browser credential
and history theft, the downloading and execution of additional malware payloads,
screen captures and keylogging, as well as file and system management
capabilities. Both NetWire and Remcos are commercial RATs that are available for
online for as little as $10 per month, and both include following the
Malware-as-a-Service (MaaS) model, offering their customers subscription-based
services with choice of licensing plans, 24/7 customer support and periodic
software updates.
Digital transformation: 4 strategy questions to ask
Internal buy-in is the most important aspect of any digital transformation and
adoption strategy, and the easiest way to help promote that is to identify
internal champions. Clearly defining the team responsible for the
implementation of a new tool or process will help give an incentive for that
team to ensure adoption is prioritized throughout the organization. It will
also help clarify where employees can direct questions. ... Training teams on
new processes and tools is easier said than done. It’s important to find a
better way to train, not just to ensure that digital transformation is
successful and to make sure training really sinks in, but also to make sure
your employees feel supported. Building effective training programs is a great
way to show your employees that you’re invested in their success and their
careers more broadly – helping to increase retention. ... Finally, be sure to
set measurable, attainable goals around your digital transformation strategy.
These may look vastly different from tool to tool or organization to
organization, but adoption will increase if every user understands how
transformation efforts will be evaluated.
5 Ways Machine Learning Is Revolutionizing the Healthcare Industry
Machine learning established new methods in drug discovery, such as precision
medicine and next-generation sequencing, which can ensure a drug has the
intended effect on patients. With the implementation of machine learning
techniques, medical experts can develop algorithms to treat disease
progression and design specific treatments for each patient, like those with
Type 2 diabetes. ... Machine learning aids medical experts in determining the
risk for each patient, depending on their symptoms, past medical records, and
family history. ML streamlines the process of finding treatments for evolving
illnesses, as well as helping researchers to track possible pandemics and to
understand better why some diseases are more prevalent in specific cultures
and demographics. ... Modern hospitals are high-tech environments run by
advanced machines and the staff who are trained. The hospitals are
increasingly shifting towards automation, to a future where diagnoses can be
made accurately. Machine learning can accelerate disease diagnostics and make
the risk of misdiagnosis less likely.
Apps that help parents protect kids from cybercrime may be unsafe too
Parental control apps need many permissions to access particular systems and
functions on devices. 80% of parental control apps request access to location,
contacts and storage. While these permissions help the apps carry out detailed
monitoring, some of them may not be necessary for the app to function as
described. For instance, several apps designed to monitor children’s online
activity ask for permissions such as “read calendar”, “read contacts” and
“record audio” — none of which are justified in the app description or the
privacy policy. Many are considered “dangerous permissions”, which means they
are used to access information that could affect the user’s privacy and make
their device more vulnerable to attack. For example, Boomerang requests more
than 91 permissions, 16 of which are considered “dangerous”. The permission
“access fine location” for instance, allows the app to access the precise
geographic location of the user. The “read phone state” allows the app to know
your phone number, network information and status of outgoing calls. It’s not
just the apps that get that information. Many of these apps embed data hungry
third-party software development kits (SDKs). SDKs are a set of software tools
and programs used by developers to save them from tedious coding.
Company uses cognitive neuroscience to help train police officers
The brain has two systems, what you're consciously aware of, and then the
non-conscious part, where almost everything happens. That's where your drives
and biases and urges and impulses all come up from this, what we call the
backchannel of the brain, the non-conscious part of the brain. Most of our
actions and behaviors are initiated there a lot of times without our
awareness. When we are thinking through something and thinking through answers
or trying to problem-solve, we can direct our conscious brain to kind of
override some of those impulses and urges and really take control of what
we're doing and what we're thinking and how we're behaving. But when we're
under stress, our brain is built for the non-conscious brain to take over, to
help us with survival, or to help us get out of a scrape, something like that.
With police officers, when they've been trained how to respond in certain ways
to help people out, to de-escalate events, things like that, when stress
starts to rise and get higher, then their non-conscious brain really starts to
take over the processing. And that's where even really good officers can do
things they later regret is because the non-conscious brain and those urges
and impulses to say something or do something happens, and they might regret
that later.
How 4 cities are modernizing their IT infrastructure through the cloud
The city's cybersecurity team leads threat management and operates a 24-hour
security operations center. The team works with more than 100 city agencies
and offices to ensure systems are built and operated in a secure manner to
make sure public assistance and healthcare are not compromised. NYC Cyber
Command also manages an NYC Secure app that alerts users to unsecure Wi-Fi
networks, unsafe Android apps and system tampering. The team uses a cloud
infrastructure to find and mitigate threats. The Cyber Command uses a variety
of Google cloud services including Cloud Storage, Computer Storage, Kubernetes
Engine and Workspace. The team uses BigQuery to analyze batch and streaming
data. When the pandemic started, DC Water already had 90% of the
organization's systems on the cloud, according to a blog post on Microsoft.
The final step was moving in-person operations and services. The organization
worked with ESRI to move applications, operational processes and customer
requests to Azure. Goals for this work included improving data security and
replacing paper processes with digital ones. Durmus Cesur, the manager of work
and asset management for DC Water, told Microsoft in the blog post that Azure
was the best solution to provide continuous availability and scalability.
Ransom Payments Have Nearly Tripled
A new report from Palo Alto Networks -- which uses data from ransomware
investigations, data-leak sites, and the Dark Web — found 337 victims in 56
industries, with manufacturing, healthcare, and construction companies
suffering 39% of ransomware attacks in 2020. In addition, ransom demands
skyrocketed during the year, doubling both the highest ransom demand — to $30
million—and the highest-known paid ransom, $10 million. The average victim
paid more than $312,000, almost a third of the average demand. ... The Palo
Alto report combines two sources of the threat intelligence: 252 incidents
investigated by the company's data-breach response service over the past two
years, and a survey of public leak sites and the Dark Web. Almost two thirds
of the incident response cases investigated by the company came in one of four
industries in 2020: healthcare, manufacturing, information technology, or
construction. ... "As organizations shifted to remote workforces due to the
COVID-19 pandemic, ransomware operators adapted their tactics accordingly,
including the use of malicious emails containing pandemic-based subjects and
even malicious mobile apps claiming to offer information about the virus," the
company stated.
Importance of Teaching Data Science in CS Programs
Besides being a lucrative career, data science is among the careers of
tomorrow. New innovations in the industrial sectors are highly reliant on
data. Technology is becoming dynamic and more data is generated as more people
access the internet. With huge amounts of data, industries rely on data
scientists to make smart business decisions. In the current digital world,
data literacy is very important. People should learn how they can generate
meaningful insights from raw data. Data is an untapped potential that can be
used to develop various sectors. Fortunately, with the inception of machine
learning technologies, organizations can predict and classify information
accurately and intelligently. Data science, machine learning, and other
similar technologies are subsets of artificial intelligence, which are the
driving force behind future products such as self-driving cars and autonomous
robots. Such developments are not fiction anymore. The emergence of
reinforcement learning and natural language processing has also contributed to
these advancements. ... The importance and urgency of data science in the 21st
Century cannot be ignored. From providing great insights, statistics, aiding
decision-making to hire suitable candidates, data science is overly valuable.
Five Steps To Thinking Like A Software Company
Leading companies feature software stacks that are modular, facilitating rapid
innovation. Their developers frequently build in-house software products or
platforms by leveraging free, but valuable, open-source software, as well as
licensed components for routine functionality. This allows them to create
applications faster. One executive stressed the importance of designing
components with change in mind, because reconfiguring is always better than
rewriting code. Another executive told me that every line of code within this
decentralized architecture has a clear owner so that there is specific
responsibility for each and every software component. To be clear, commercial
solutions have an important role to play and should be a part of the software
stack. But it’s the own-account software that matters most. ... In contrast,
firms that lead with code typically begin by aiming to solve a focused business
problem. They build and iterate on new features and products. Executives at
these firms told me that until you try something out and see how your customers,
suppliers, or employees react, and whether your business improves as a result,
you can’t be sure of what to build.
Quote for the day:
"If you don't start somewhere, you're
gonna go nowhere." -- Bob Marley
No comments:
Post a Comment