Making computers reason and learn by analogy

Structure-mapping engine enables computers to reason and learn like humans, including solving moral dilemmas

Credit: © rolffimages / Fotolia

Northwestern University's Ken Forbus is closing the gap between humans and machines.

Using cognitive science theories, Forbus and his collaborators have developed a model that could give computers the ability to reason more like humans and even make moral decisions. Called the structure-mapping engine (SME), the new model is capable of analogical problem solving, including capturing the way humans spontaneously use analogies between situations to solve moral dilemmas.

"In terms of thinking like humans, analogies are where it's at," said Forbus, Walter P. Murphy Professor of Electrical Engineering and Computer Science in Northwestern's McCormick School of Engineering. "Humans use relational statements fluidly to describe things, solve problems, indicate causality, and weigh moral dilemmas."

The theory underlying the model is psychologist Dedre Gentner's structure-mapping theory of analogy and similarity, which has been used to explain and predict many psychology phenomena. Structure-mapping argues that analogy and similarity involve comparisons between relational representations, which connect entities and ideas, for example, that a clock is above a door or that pressure differences cause water to flow.

Analogies can be complex (electricity flows like water) or simple (his new cell phone is very similar to his old phone). Previous models of analogy, including prior versions of SME, have not been able to scale to the size of representations that people tend to use. Forbus's new version of SME can handle the size and complexity of relational representations that are needed for visual reasoning, cracking textbook problems, and solving moral dilemmas.

"Relational ability is the key to higher-order cognition," said Gentner, Alice Gabrielle Twight Professor in Northwestern's Weinberg College of Arts and Sciences. "Although we share this ability with a few other species, humans greatly exceed other species in ability to represent and reason with relations."

Supported by the Office of Naval Research, Defense Advanced Research Projects Agency, and Air Force Office of Scientific Research, Forbus and Gentner's research is described in the June 20 issue of the journal Cognitive Science. Andrew Lovett, a postdoctoral fellow in Gentner's laboratory, and Ronald Ferguson, a PhD graduate from Forbus's laboratory, also authored the paper.

Many artificial intelligence systems -- like Google's AlphaGo -- rely on deep learning, a process in which a computer learns examining massive amounts of data. By contrast, people -- and SME-based systems -- often learn successfully from far fewer examples. In moral decision-making, for example, a handful of stories suffices to enable an SME-based system to learn to make decisions as people do in psychological experiments.

"Given a new situation, the machine will try to retrieve one of its prior stories, looking for analogous sacred values, and decide accordingly," Forbus said.

SME has also been used to learn to solve physics problems from the Advanced Placement test, with a program being trained and tested by the Educational Testing Service. As further demonstration of the flexibility of SME, it also has been used to model multiple visual problem-solving tasks.

To encourage research on analogy, Forbus's team is releasing the SME source code and a 5,000-example corpus, which includes comparisons drawn from visual problem solving, textbook problem solving, and moral decision making.

The range of tasks successfully tackled by SME-based systems suggests that analogy might lead to a new technology for artificial intelligence systems as well as a deeper understanding of human cognition. For example, using analogy to build models by refining stories from multiple cultures that encode their moral beliefs could provide new tools for social science. Analogy-based artificial intelligence techniques could be valuable across a range of applications, including security, health care, and education.

"SME is already being used in educational software, providing feedback to students by comparing their work with a teacher's solution," Forbus said. But there is a vast untapped potential for building software tutors that use analogy to help students learn."

---

Story Source:

The above post is reprinted from materials provided by Northwestern University. Note: Materials may be edited for content and length.

---

Journal Reference:

1. Kenneth D. Forbus, Ronald W. Ferguson, Andrew Lovett, Dedre Gentner. Extending SME to Handle Large-Scale Cognitive Modeling. Cognitive Science, 2016; DOI:10.1111/cogs.12377

Palaeontologist discovers new species of 200 million-year-old 'British' marine reptile

A new type of ichthyosaur, an extinct marine reptile alive at the same time as the dinosaurs, has been identified by a palaeontologist from a fossil found in an old quarry in Nottinghamshire.

A new type of ichthyosaur, an extinct marine reptile alive at the same time as the dinosaurs, has been identified by a Manchester palaeontologist from a fossil found in an old quarry in Nottinghamshire.

Similar-shaped to dolphins and sharks, ichthyosaurs -- often misidentified as 'swimming dinosaurs' -- swam the seas of Earth for millions of years during the Triassic, Jurassic and Cretaceous periods. The Nottinghamshire fossil is from the earliest part of the Jurassic Period -- 200 million years ago -- and only a handful of ichthyosaur species are known from this period, making the discovery very significant. It is also the first time a species of this geological age has been found outside of Dorset and Somerset.

Dean Lomax, a Palaeontologist and Honorary Scientist at The University of Manchester, examined the specimen after seeing it on a visit to Leicester's New Walk Museum, which acquired the fossil in 1951, and spotted some unusual features. The specimen is relatively complete, consisting of a partial skeleton including a skull, pectoral bones, limbs, pelvis bones, ribs and vertebrae. However, the bones are disorderly -- it appears that the carcass 'nosedived' into the seabed before it became fossilised, which may have restricted previous study.

"When I first saw this specimen, I knew it was unusual," said Dean. "It displays features in the bones -- especially in the coracoid (part of the pectoral girdle) -- that I had not seen before in Jurassic ichthyosaurs anywhere in the world. The specimen had never been published, so this rather unusual individual had been awaiting detailed examination."

Dr Mark Evans, Palaeontologist and Curator of Natural Sciences at New Walk Museum, said: "Parts of the skeleton had previously been on long-term loan to ichthyosaur specialist and former museum curator Dr Robert Appleby, and had only returned to the museum in 2004 after he sadly passed away. He was clearly intrigued by the specimen, and although he worked on it for many years, he had identified it as a previously known species but never published his findings."

Dean has named the new species Wahlisaurus massarae in honour of two palaeontologists (Professor Judy Massare and Bill Wahl) who have contributed significantly to the study of ichthyosaurs, and who first introduced Dean to studying them.

"Both Judy and Bill have been tremendous mentors for me. They have significantly contributed to palaeontology, especially the study of ichthyosaurs, and I cannot think of a better way to remember them by naming this new ichthyosaur in their honour. Their names will be set in stone forever, pun intended!"

The specimen is the first new genus of ichthyosaur from the British Early Jurassic to be described since 1986. Thousands of specimens from this time are known, and many of these have been examined -- and continue to be re-examined in light of new knowledge and technologies. However, as the specimen is from a practically unknown location for the discovery of ichthyosaurs, any new discovery could be of real scientific significance. This new species is also important for our understanding of ichthyosaur species diversity, and their geographical distribution during the Early Jurassic.

---

Story Source:

The above post is reprinted from materials provided by Taylor & Francis. Note: Materials may be edited for content and length.

---

Journal Reference:

1. Dean R. Lomax. A new leptonectid ichthyosaur from the Lower Jurassic (Hettangian) of Nottinghamshire, England, UK, and the taxonomic usefulness of the ichthyosaurian coracoid. Journal of Systematic Palaeontology, 2016; 1 DOI:10.1080/14772019.2016.1183149

iPhone Maker Foxconn Cuts 60,000 Factory Jobs and Replaces Them With Robots

Foxconn has reduced 60,000 employees as the company has replaced robots in place of the workers. Earlier, Foxconn had 110,000 employees, now the number has reached 50,000. Next time when you are going to buy an iPhone, it is likely to have been created by a robot.

iPhone Maker Foxconn Cuts 60,000 Factory Jobs and Replaces Them With Robots

The largest contact electronics maker of the world, Foxconn announced that it has automated the place of 60,000 factory worker with robots, reports BBC. The job cuts are probably taken to replace humans with robots for many of the manufacturing tasks associated with our operations” with robots, according to the company.

Foxconn manufactures Apple’s devices like iPhone and iPad, Samsung’s Galaxy Series devices, Sony’s Playstation 4 including other devices from the reputed tech brands.

“We are applying robotics engineering and other innovative manufacturing technologies to replace repetitive tasks previously done by employees, and through training, also enable our employees to focus on higher value-added elements in the manufacturing process, such as research and development, process control and quality control,” according to the statement.

We will continue to harness automation and manpower in our manufacturing operations, and we expect to maintain our significant workforce in China.”

“The Foxconn factory has reduced its employee strength from 110,000 to 50,000, thanks to the introduction of robots. It has tasted success in reduction of labour costs.”

It seems that the major tech companies are adopting robots instead of labor force. The debate regarding the adoption of robots is making headlines in the US with the former CEO of McDonalds, En Rensi saying “It’s cheaper to buy a $35,000 robotic arm than it is to hire an employee who’s inefficient making $15 an hour bagging french fries.”

Foxconn has taking steps to replace labourers with robots for a long time. It is likely that the company is taking this step to increase profit and save money.

No doubt, Robots are going to make our work more easier and faster, but at the same time it can give rise to unemployment crisis across the world. Robots could leave more than half of the population Jobless in the next 30 years, according to a computer scientist who said that Artificial Intelligence’s threat to the economy should not be underestimated

Source : TechViral

How to Make E-Books with Google Docs

While you are making your documents on either Microsoft Word or on Google Docs that both are most popular software to work on your documents, you might want to turn up your work to PDF file for any reasons. This can be important for some users also to transform their creations into PDF files. To convert your documents into PDF files you will have to go through some method so that you can ultimately change your Google Docs documents into PDF files. Here in this article we have discussed about the method for converting or making eBooks with Google Docs, just go through it as it is given below.

How to Make E-Books with Google Docs

The method is quite simple and easy and you just need to follow some simple straight guide that will help you to easily convert the files, So follow some simple steps below to proceed.

Steps To Make eBooks with Google Docs:

1. First of all make your documents on either Google Docs or Microsoft Word and then save them on your local Drive.

2. After that Open up your Google Drive account i.e cloud storage account if you have it and if you had not created it still then you have to first create it and then open it up on your device.

3. Now find the file for the documents that you had created and saved before, and then upload it to the Google Drive. This should be done correctly unless your file would get corrupted and cannot be accessed later on.

4. At last, you have to download the same file that you have uploaded on the Google Drive but before downloading you will have to set the file format to the ePub, eBook or PDF. You can do this easily from the options available on the Google Drive.

5.Your document files will be then downloaded as a PDF file, ePub file etc. You can access then as other common files with such formats.

If you don’t want to go for this process then there are some applications for PC like Adobe Stanza and Calibre which you can use to convert your files to the different formats that include PDF, ePUB, ebooks etc.

So above guide is all about Make eBooks with Google Docs. Making or transforming your work on Google Docs into PDF files was not that easy as you have found it on this article. You can now easily make out the PDF files from your Google Docs projects and then view them with any suitable PDF reader application either on your computer or on your smartphone. Just follow up the method given above and try it yourself to create your own eBooks with Google Docs.

Source: TechViral

New study finds major earthquake threat from the Riasi fault in the Himalayas

New geologic mapping in the Himalayan mountains of Kashmir between Pakistan and India suggests that the region is ripe for a major earthquake that could endanger the lives of as many as a million people.

New geologic mapping in the Himalayan mountains of Kashmir between Pakistan and India suggests that the region is ripe for a major earthquake that could endanger the lives of as many as a million people.

Scientists have known about the Riasi fault in Indian Kashmir, but it wasn't thought to be as much as a threat as other, more active fault systems. However, following a magnitude 7.6 earthquake in 2005 on the nearby Balakot-Bagh fault in the Pakistan side of Kashmir -- which was not considered particularly dangerous because it wasn't on the plate boundary -- researchers began scrutinizing other fault systems in the region.

What they found is that the Riasi fault has been building up pressure for some time, suggesting that when it does release or "slip," the resulting earthquake may be large -- as much as magnitude 8.0 or greater.

Results of the new study, which was funded by the National Science Foundation, have been accepted for publication by theGeological Society of America Bulletin, and published online.

"What we set out to learn was how much the fault has moved in the last tens of thousands of years, when it moved, and how different segments of the fault move," said Yann Gavillot, lead author on the study who did much of the work as a doctoral student at Oregon State University. "What we found was that the Riasi fault is one of the main active faults in Kashmir, but there is a lack of earthquakes in the more recent geologic record.

"The fault hasn't slipped for a long time, which means the potential for a large earthquake is strong. It's not a question of if it's going to happen. It's a matter of when."

There is direct evidence of some seismic activity on the fault, where the researchers could see displacement of Earth where an earthquake lifted one section of the fault five or more meters -- possibly about 4,000 years ago. Written records from local monasteries refer to strong ground-shaking over the past several thousand years.

But the researchers don't have much evidence as to how frequent major earthquakes occur on the fault, or when it may happen again.

"The Riasi fault isn't prominent on hazard maps for earthquake activity, but those maps are usually based more on the history of seismic activity rather than the potential for future events," said Andrew Meigs, a geology professor in OSU's College of Earth, Ocean, and Atmospheric Sciences and co-author on the study. "In actuality, the lack of major earthquakes heightens the likelihood that seismic risk is high."

The researchers say 50 percent of the seismic "budget" for the fault can be accounted for with the new information. The budget is determined over geologic time by the movement of the tectonic plates. In that region, the India tectonic plate is being subducted beneath the Asia plate at a rate of 14 millimeters a year; the Riasi fault accounts for half of that but has no records of major earthquakes since about 4,000 years ago, indicating a major slip, and earthquake, is due.

"In the last 4,000 years, there has only been one major event on the Riasi fault, so there is considerable slip deficit," Meigs said. "When there is a long gap in earthquakes, they have the potential to be bigger unless earthquakes on other faults release the pressure valve. We haven't seen that. By comparison, there have been about 16 earthquakes in the past 4,000 years in the Cascadia Subduction Zone off the Northwest coast of the United States."

Gavillot said a major earthquake at the Riasi fault could have a major impact on Jammu, the Indian capital of the Indian state of Jammu and Kashmir, which has a population of about 1.5 million people. Another 700,000 people live in towns located right on the fault.

"There are also several dams on the Chenab River near the fault, and a major railroad that goes through or over dozens of tunnels, overpasses and bridges," Gavillot said. "The potential for destruction is much greater than the 2005 earthquake."

The 2005 Kashmir earthquake killed about 80,000 people in Pakistan and India.

---

Story Source:

The above post is reprinted from materials provided by Oregon State University. Note: Materials may be edited for content and length.

---

Journal Reference:

1. Y. Gavillot, A. Meigs, D. Yule, R. Heermance, T. Rittenour, C. Madugo, M. Malik. Shortening rate and Holocene surface rupture on the Riasi fault system in the Kashmir Himalaya: Active thrusting within the Northwest Himalayan orogenic wedge. Geological Society of America Bulletin, 2016; B31281.1 DOI:10.1130/B31281.1

How comets break up, make up

A new study indicates the bodies of some periodic comets -- objects that orbit the sun in less than 200 years -- may regularly split in two, then reunite down the road.

For some comets, breaking up is not that hard to do.

A new study led by Purdue University and the University of Colorado Boulder indicates the bodies of some periodic comets -- objects that orbit the sun in less than 200 years -- may regularly split in two, then reunite down the road.

In fact, this may be a repeating process fundamental to comet evolution, according to the study, which is being published inNature on June 1.

The team, led by Purdue postdoctoral fellow Masatoshi Hirabayashi and CU-Boulder Distinguished Professor Daniel Scheeres, studied several comets, primarily a bizarre rubber duck-shaped object known as 67P/Churyumov-Gerasimenko (67P). Images of 67P show two cracks, each longer than an American football field, on the comet's neck that connects its two larger lobes.

In order to reconstruct the past life of 67P, the team used numerical models in which the spin rate was cranked up from its roughly one rotation every 12 hours today to one rotation every 7 to 9 hours. The models showed the faster spin would lead to more stress and the formation of two similar cracks on the neck of 67P in the same location.

"Our spin analysis predicted exactly where these cracks would form," said Scheeres of CU-Boulder's aerospace engineering sciences department. "We now have a new understanding of how some comets may evolve over time."

Often referred to as "dirty snowballs," comets are made of ice, rocks and dust. Comet 67P is "bilobed" meaning it has two larger parts connected by a thinner neck.

Scheeres said there are several factors that can cause comet nuclei to spin faster. During flybys of the sun or Jupiter, for example, periodic comets like 67P can get torqued by gravity, causing them to either spin up or spin down. The spin also can be affected by periodic comet "outgassing," when icy compounds like carbon dioxide and ammonia shift directly from a frozen state to gaseous state and blow off the surface.

The models run by the team showed that if 67P's spin is increased to less than seven hours per rotation, the head will pop off, said Scheeres. So what happens then?

"The head and body aren't going to be able to escape from each other," he said. "They will begin orbiting each other, and in weeks, days or even hours they will come together again during a slow collision, creating a new comet nucleus configuration."

This pattern could go on for the life of the comet, said Scheeres.

Bilobed comets may turn out to be fairly common. Of the seven comets that have been imaged in high resolution by astronomers, five of those -- including P67 and Comet Halley -- are bilobed, said Scheeres. Studies of the bilobed comets by the team indicate they all are similar in their volume ratios between each lobe, meaning they probably go through the same break-up/make-up cycles as 67P.

Discovered in 1969 and visited by the European Space Agency's Rosetta spacecraft in 2014, 67P is roughly 2.5 miles on a side and orbits the sun every 6.5 years. The team showed that the comet's spin rate can change chaotically, driven by outgassing events and its changing orbit driven by flybys of Jupiter.

To show how this comet-sun interaction affected the past evolution of 67P's spin period, the researchers numerically modeled 1,000 comet "clones" of 67P under varying conditions going back 5,000 years. Five thousand years was selected because it is the approximate lifetime of a "Jupiter family comet" like 67P, whose orbit is affected by the gravity of not only the sun but the gas giant Jupiter, the largest planet in our solar system, said Scheeres.

Periodic comets like 67P are thought to originate in the Kuiper Belt, a vast region beyond Neptune's orbit harboring billions of comets and icy moons. The team hypothesized that the repeated break-up and make-up of bilobed comets may have caused them to erode too much to have survived their journeys into the inner solar system 4 billion years ago when it was a shooting gallery of asteroids, moons and protoplanets.

Other study authors include Assistant Professor Jay McMahon of CU-Boulder, Steven Chesley of the Jet Propulsion Laboratory in Pasadena and Simone Marchi of the Southwest Research Institute Planetary Science Directorate in Boulder. Purdue's Hirabayashi received his doctorate at CU-Boulder under Scheeres in 2015.

Scheeres is the radio science team leader for NASA's OSIRIS-Rex mission, now slated to launch in September from Florida to visit the near-Earth asteroid, Bennu.

---

Story Source:

The above post is reprinted from materials provided by University of Colorado at Boulder. Note: Materials may be edited for content and length.

---

Journal Reference:

1. Masatoshi Hirabayashi, Daniel J. Scheeres, Steven R. Chesley, Simone Marchi, Jay W. McMahon, Jordan Steckloff, Stefano Mottola, Shantanu P. Naidu, Timothy Bowling. Fission and reconfiguration of bilobate comets as revealed by 67P/Churyumov–Gerasimenko.Nature, 2016; DOI: 10.1038/nature17670

Universe's first life might have been born on carbon planets

Our Earth consists of silicate rocks and an iron core with a thin veneer of water and life. But the first potentially habitable worlds to form might have been very different. New research suggests that planet formation in the early universe might have created carbon planets consisting of graphite, carbides, and diamond. Astronomers might find these diamond worlds by searching a rare class of stars.

Our Earth consists of silicate rocks and an iron core with a thin veneer of water and life. But the first potentially habitable worlds to form might have been very different. New research suggests that planet formation in the early universe might have created carbon planets consisting of graphite, carbides, and diamond. Astronomers might find these diamond worlds by searching a rare class of stars.

"This work shows that even stars with a tiny fraction of the carbon in our solar system can host planets," says lead author and Harvard University graduate student Natalie Mashian.

"We have good reason to believe that alien life will be carbon-based, like life on Earth, so this also bodes well for the possibility of life in the early universe," she adds.

The primordial universe consisted mostly of hydrogen and helium, and lacked chemical elements like carbon and oxygen necessary for life as we know it. Only after the first stars exploded as supernovae and seeded the second generation did planet formation and life become possible.

Mashian and her PhD thesis advisor Avi Loeb (Harvard-Smithsonian Center for Astrophysics) examined a particular class of old stars known as carbon-enhanced metal-poor stars, or CEMP stars. These anemic stars contain only one hundred-thousandth as much iron as our Sun, meaning they formed before interstellar space had been widely seeded with heavy elements.

"These stars are fossils from the young universe," explains Loeb. "By studying them, we can look at how planets, and possibly life in the universe, got started."

Although lacking in iron and other heavy elements compared to our Sun, CEMP stars have more carbon than would be expected given their age. This relative abundance would influence planet formation as fluffy carbon dust grains clump together to form tar-black worlds.

From a distance, these carbon planets would be difficult to tell apart from more Earth-like worlds. Their masses and physical sizes would be similar. Astronomers would have to examine their atmospheres for signs of their true nature. Gases like carbon monoxide and methane would envelop these unusual worlds.

Mashian and Loeb argue that a dedicated search for planets around CEMP stars can be done using the transit technique. "This is a practical method for finding out how early planets may have formed in the infant universe," says Loeb.

"We'll never know if they exist unless we look," adds Mashian.

---

Story Source:

The above post is reprinted from materials provided by Harvard-Smithsonian Center for Astrophysics. Note: Materials may be edited for content and length.

---

Journal Reference:

1. Natalie Mashian, Abraham Loeb. CEMP stars: possible hosts to carbon planets in the early universe. Monthly Notices of the Royal Astronomical Society, 2016