close

micrsoft

micrsoft

Microsoft Gives Productivity Tools More AI Chops

xl-2017-microsoft-ai-1

Microsoft on Wednesday announced new artificial intelligence features and functionality for several of its flagship products and services, including Office 365, Cortana and Bing, at an event in San Francisco. Harry Shum, EVP of Microsoft Artificial Intelligence and Research, demonstrated some of the new capabilities

Building on the progress the company has made in integrating AI over the past year, the new enhancements are designed to help users perform increasingly complex and complicated tasks.

“AI has come a long way in the ability to find information, but making sense of that information is the real challenge,” said Kristina Behr, a partner design and planning program manager with Microsoft’s Artificial Intelligence and Research group.

One of the advances, machine reading comprehension, will improve an AI-based system’s understanding of context — for example, recognizing that one’s cousin is a family member.

Search Benefits

Bing users will get more personalized answers, Microsoft said, such as restaurant recommendations based on travel destinations, or a greater variety of answers to offer different perspectives on a topic.

AI guidance might help users figure out the exact questions they want to ask if their search queries are too vague to deliver meaningful results.

Bing will utilize visual search as a way to help people seek information about items or objects in photos. This functionality will comprise the use of object recognition along with machine reading comprehension.

An Office AI

Microsoft announced improvements in AI integration with its Office 365 tools as well. The spreadsheet program Excel, for example, will employ machine learning as a way to analyze data and predict trends via pivot tables and charts. The new AI tools also will be able to extract insights from smaller sets of data.

Importantly, use of the new AI functionality won’t require more advanced training in Excel for users to determine which sets of data should be used to gain deeper insights.

More Conscious Cortana

Microsoft’s virtual assistant Cortana will get an upgrade as well, allowing it to make use of machine reading comprehension to summarize search results. Cortana also will be able to sort through a user’s emails to identify the most important ones and even read them aloud during a commute, for example. Use of that functionality will extend to emails from multiple accounts, including on competing services such as Gmail.

Cortana also has been enhanced with “skills chaining,” allowing it to suggest additions to one’s calendar after tickets have been booked, for example.

Astute AI

These AI advances likely will be subtle and may go unnoticed by many users.

“AI will creep more and more into our lives,” said Roger Entner, principal analyst at Recon Analytics.

“Ideally it will help to make tools like Bing, Cortana and Office 365 more userful,” he told TechNewsWorld.

“This is the good AI, not the Terminator AI,” noted Jim McGregor, principal analyst at Tirias Research.

“In most cases, AI functions will be transparent to users just like they are in search engines today,” he told TechNewsWorld.

The AI functionality simply makes the tools more efficient and useful.

“Think of this as reducing some of the simple tasks — not the human part of the equation,” McGregor suggested, “and think about intelligent digital assistants and word predictors or spell checkers that are more reliable than the ones we have today. This will be the most common form of AI — embedded solutions that make the tools we use better, from search engines to scientific research.”

How Far to Take It

Although the most extreme AI danger may lie in the rise of machines that overpower human controls, there are more subtle menaces to consider when it comes to the place it will take in our lives.

“It is upon us to draw a line,” said Entner.

“Examples like Google and its all-encompassing ‘data kraken’ make it clear how blurry the line is between really useful and really creepy — like when your Android phone constantly asks you if you are actually shopping at the store you are at,” he pointed out. “Big brother [could be] watching all the time

read more
GoogleIntelmicrsoft

Intel, Microsoft, Google Scramble for Solutions as Patches Slow Systems

xl-2017-female-system-engineer-1

Cybersecurity | Mobile Tech | IT Management | Cloud Services
To compete in today’s business world you need to make the most of the available technologies. ALL EC has the resources you need to help make informed buying decisions. Keep your business moving forward.
Visit ALL EC today.

Major tech companies, including Intel, Microsoft and Google, scrambled to calm the mood this week after a large number of computer users reported performance problems linked to security updates for the Spectre and Meltdown vulnerabilities.

A firestorm of criticism has erupted over the response to the chip flaws, which researchers at Google’s Project Zero discovered in 2016. Months passed before the problems were disclosed to the public. Further, the security patches released in recent days have been blamed for performance problems, including slowdowns in many systems. The fixes reportedly rendered a smaller number of systems unbootable.

Intel CEO Brian Krzanich on Thursday sent an open letter to the technology industry, pledging the company would make frequent updates and be more transparent about the process, and that it would report security issues to the public in a prompt manner.

Design Flaw

Intel Executive Vice President Navin Shenoy on Wednesday issued an update on the impact of the patches on performance, saying that eighth-generation Kaby Lake and Coffee Lake platforms would see less than a 6 percent performance decrease. However, users running Web applications with complex Javascript operations might see a 10 percent reduction.

The seventh-generation Kaby Lake platforms would experience a 7 percent reduction, and the impact on the sixth-generation Skylake platforms would be slightly higher at 8 percent.

Intel released numerous statements after the vulnerabilities were made public, and it shot down reports that its chips were the only ones at risk.

However, the Rosen Law Firm on Wednesday announced that it had filed a class action suit against Intel, alleging a failure to disclose the design flaw. The complaint cited reports that Intel had been warned of the problem. An Intel spokesperson was not immediately available to comment for this story.

Project Zero researchers discovered serious security flaws caused by “speculative execution,” a technique used by modern CPUs to optimize performance, Matt Linton, senior security engineer at Google Cloud, and Matthew O’Connor, office of the CTO, wrote in an online post.

G Suite and Google Cloud platforms have been updated to protect against known attacks, the company said, though it acknowledged concerns that a variant of Spectre is considered more difficult to defend against.

Microsoft and others in the industry were notified of the issue several months ago under a nondisclosure agreement, Terry Myerson, executive vice president of Microsoft’s Windows and Devices group, noted earlier this week in an online post. The company immediately began engineering work on updates to mitigate the risk.

The flaw could allow a nonprivileged user to access passwords or secret keys on a computer or a multitenant cloud server, explained Stratechery analyst Ben Thompson in a post Myerson referenced.

Contrary to Intel’s protests, the potential risk from Meltdown is due to a design flaw, Thompson also noted.

Users of Windows 8 or Windows 7 systems using Haswell or older CPUs and would see a decrease in system performance after patching the flaw, Myerson noted.

Apple released updates for iOS, macOS High Sierra, and Safari on Sierra and El Capitan, noting the issue relates to all modern processors and affects nearly all computers and operating systems.

However there have been no reported compromises of customer data, Apple added, and Apple Watch is not affected by Meltdown or Spectre.

Performance Over Prudence

“The Meltdown and Spectre vulnerabilities require adjustment to critical, low-level interfaces in affected operating systems,” said Mark Nunnikhoven, vice president of cloud security at Trend Micro.

“Given the scale of the issue, the patches by Microsoft, Apple, Google and others have been very successful,” he told TechNewsWorld.

Still, there have been problems in some cases, Nunnikhoven said, noting that Microsoft and AMD have been pointing fingers at one another following reports of computers slowing down or in some cases not booting.

Microsoft has suspended automatic updates and is working with AMD on a solution, it said in a security bulletin.

Like most organizations, chip manufacturers long have prioritized speed over security,” said Ryan Kalember, senior vice president of cybersecurity strategy atProofpoint, “and that has led to a tremendous amount of sensitive data being placed at risk of unauthorized access via Meltdown and Spectre.

The software patch required to fix Meltdown can slow computer processors down by as much as 30 percent, said Alton Kizziah, vice president of global managed services at Kudelski Security.

“Organizations need to test patches before installing them to make sure that systems that may already be pushed to their limits won’t crash and cease functioning as a result of the patch,” he told TechNewsWorld. Also, those using Microsoft patches may need to make adjustments to their registry keys to avoid interference with antivirus software.

read more