Blog: AI in Libraries

Like to be involved.

Sunday, November 28, 2021

New directions in AI: formation of an IFLA Special Interest Group on Artificial Intelligence

There is an online meeting on New directions in AI: formation of an IFLA Special Interest Group on Artificial Intelligence on 6 December 2021 at 4pm UTC (UK time); 5pm CET; 11am US EST. This exploratory meeting will: "give an overview of the current state of AI in libraries; discuss the goals and objectives; gather 25 signatories who intend to actively participate in the activities of the SIG for a petition to be submitted to the Professional Council; propose a satellite meeting and main session at IFLA WLIC 2022 in Dublin, Ireland." "Artificial intelligence applications are increasingly a part of the library space: in chatbots, embedded in library systems, used for automated indexing and classification, and integral to robots. The IT Section is sponsoring the formation of a Special Interest Group in AI (AI SIG). ... If the SIG is approved we will also hold the first business meeting to nominate a Convenor and seek volunteers to serve in roles including Secretary and Communications Coordinator. Registration at 

Posted by Sheila Webber at 16:07   ...... ' 

Kaolin: A Library for 3D Intelligence

Has to be useful for building and testing 3D intelligence.

Kaolin: The first comprehensive library for 3-D deep learning research  by Ingrid Fadelli , Tech Xplore

As most real-world environments are three-dimensional, deep learning models designed to analyze videos or complete tasks in real-world environments should ideally be trained on 3-D data. Technological tools such as robots, self-driving vehicles, smartphones, and other devices are currently generating a growing amount of 3-D data that could eventually be processed by deep learning algorithms.

Up until now, however, training deep learning algorithms on this vast amount of 3-D data has been relatively difficult, as the necessary tools and platforms are only accessible to some artificial intelligence (AI) researchers. To address this lack of readily available tools, a team of researchers at NVIDIA has recently created Kaolin, a PyTorch open-source library aimed at advancing and facilitating 3-D deep learning research.

"Currently, there is not a single open-source software library that supports multiple representations of 3-D data, multiple tasks, and evaluation criteria," Krishna Murthy Jatavallabhula, one of the researchers who carried out the study, told TechXplore. "We decided to address this gap in the literature by creating Kaolin, the first comprehensive 3-D deep learning library."

Kaolin, the PyTorch library presented by Jatavallabhula and his colleagues, contains a variety of tools for constructing deep learning architectures that can analyze 3-D data, which are both efficient and easy to use. It also allows researchers to load, preprocess, and manipulate 3-D data before it is used to train deep learning algorithms.  .... "

Curate Libraries for Truth?

Internet pioneer dicusses truth and knowledge on the Internet.  Mentions the rare word 'Internaut'.

Libraries Considered Hazardous   By Vinton G. Cerf
Communications of the ACM, February 2019, Vol. 62 No. 2, Page 5
10.1145/3302508

Do you remember the story of the room full of immortal monkeys typing on typewriters forever? Eventually they would produce all works ever written and that would ever be written. They would capture all truth but also everything that is false or only partly true. Were we to walk into such a place we would be confronted with an ultimate challenge: How to tell that which was true from everything else in this ultimate library?

In some ways, the contents of the Internet and especially the World Wide Web pose a similar challenge. About half the world's population is now online according to estimates by the International Telecommunication Union.a These approximately 3.8 billion people produce enormous quantities of information on Web pages, in databases, in social media, and other online platforms. While I do not mean to suggest these Internauts are no better than monkeys typing at random, there is a great deal of misinformation mixed in with very high-quality content. Some of that misinformation is a consequence of ignorance, but some is deliberately produced disinformation intended to confuse or to bend public opinion to achieve questionable ends. Ironically, some of the best quality, highly endorsed information is also wrong, not out of malevolent intent, but because it has been invalidated by the scientific method: theory, experiment, and measurement leading to proof or refutation.

If we are honest with ourselves, science is, at best, an approximation of reality. Even when they are not quite right, some theories can still be very useful. Newton's laws are useful for many computations but under conditions of acceleration, high-speed or intense gravity, one needs Einstein's refinements. And when we get to the ultra-small, we must move to quantum theory, but it doesn't account for gravity! The challenge for us is to know under what conditions the approximations are applicable.

How does all this apply to libraries? Libraries are organized accumulations of information. I almost wrote "knowledge" but that term seems to connote "truth" and we know now that all information is not true. As we accumulate more and more information, how can we curate this content so as to correctly distinguish truth from fiction? How do we cope with the discovery that what we thought was true is, in fact, false in the light of new information? Librarians have a role to play here as keepers of knowledge, but even they cannot be expected to be omniscient. What about digital content? What about online content? Can the curators of knowledge use online digital libraries to maintain and curate content, helping the users of the library to find truth and reject fiction (except, perhaps, when looking for entertainment)?    ...  "

Top 20 Python Data Science Libraries from DSC

Extensive sets with good descriptions. 

Top 20 Python libraries for data science in 2018

Posted by Igor Bobriakov

Python continues to take leading positions in solving data science tasks and challenges. Last year we made a blog post overviewing the Python’s libraries that proved to be the most helpful at that moment. This year, we expanded our list with new libraries and gave a fresh look to the ones we already talked about, focusing on the updates that have been made during the year.

Our selection actually contains more than 20 libraries, as some of them are alternatives to each other and solve the same problem. Therefore we have grouped them as it's difficult to distinguish one particular leader at the moment. .... " 

Library to Preserve Sounds

Usually don't think of sounds in general,  as opposed to music, but interesting to explore. The link has the ability to see a number of examples.

British Library exhibit to highlight the sounds it’s fighting to save
"Listen: 140 Years of Recorded Sound" runs from October 6th to March 11th next year.

By Jamie Rigg, @jmerigg

Business Leadership Mind Maps

The Bigger Plate site has a broad library of mind maps, both complete and in useful template form.  Most recently a set of business leadership mind maps, I will take a further look.  ( Apparently now requires their mind mapping software.  If this is not the case please let me know and I will update and re-post. )

On Technology and Library Users

A view of Pew research on the users of (physical) libraries. And a quiz to show if you are a library lover.  I still admit to being one.  Interesting findings are buried in the statistics.

Shopping Image Recognition by Amazon: Flow

Have spent some time on the process of shopping using in-aisle using scanning methods.  Finally got a chance to look at this method more closely using an IOS installed Amazon App: Flow.  The approach, a little mysteriously, is called Flow.  I imagine implying how it could improve the flow of the in-store shopping experience.  Here is Amazon's description.  The camera is used as a means of reading symbol codes and images.   See my previous notes on this.

I am working with a used book selling operation, and its useful to be able to quickly scan and recognize inventory.  Some time ago the process moved to scanning book barcodes and getting meta information from the Amazon database.

This test went against a number of books, variously oriented, without searching for a barcode first.   The system works quite well, faster than a barcode read.  Reading from the spine side of a book was only sometimes successful.

Overall well done, can improve interaction in store and warehouse.  Woould be better yet if it could read spine directed books.   We were also able to identify several items in a single view,  but that is not currently an operational need.

Where Does the Internet End and the Library Begin?

Interesting panel conversation.  A question that continues to evolve with the introduction of mobile abilities.  The panel talks about formal libraries.  But I think a strong additional argument is about the democratization of knowledge. Available anytime, anywhere and up to date.  Based on what we have seen, this will evolve further.

National Digital Public Library

Just launched.  A new open source library of US resources has launched in Beta.  The DPLA.  Some have been talking about how decrepit US copyright laws will influence this effort.  " ... Two years in the making, the DLPA will make available to the public 2.4 million records at its launch, including electronic images, video and audio from America's libraries, archives and museums. It also makes many scientific records available.... "  On twitter as @DPLA

Data Mining in R: Introductory Resources

The DataMining Collective sends along links to Data Mining in R.  Fairly well done as an introduction, of both the power and complexity of using the Package R.   Includes slides and audio.  A  50 minute  description of R as groups of algorithms, community and interface.  These are libraries of functions,  most useful for scientists who are used to coding, and not in general end users.    See also the presenter David Smith's blog which covers R, Data mining and related topics.

Georgia R School Opens

AnalyticsBridge posts the particulars about the online 'Georgia R School', available for as little as $110.  The R stat and visualization package is becoming increasingly popular for applications, many students are coming out of school knowing it, and many new methods are being built into its libraries.  But it is tricky enough to need some up front training for those that did not get it in school.   The curriculum here looks to be very good for starting out with some statistics background.    I also suggest joining Analyticsbridge, which often has gems like this.  I sometimes pass them along there.  ...  Plus, I note, lots more course ware is available in the coming year.

College Libraries Exploring Pinterest

My own involvement with libraries would not make me expect this.  But libraries are doing more experimentation than they used to. Moving from text centric to image focused.    I have wandered the stacks checking out the spines.  And libraries discovered long ago you could arrange books on a shelf to 'sell' them with their covers.  Now I want to see books arranged as a curriculum  ... or maybe even an extended story?  And why not do the same thing virtually?

Innovation for the Library

In the HBR:   If is clearly a challenge for the library to innovate in a way it will be able to survive.  " ... The dominance of radio, TV, and Hollywood threatened libraries with irrelevance, and now the rise of the internet, smart phones, and ebooks, threaten them with eclipse ....  So libraries are struggling. What can they do in the face of skeptics who say, "Print! Why, in God's name, do we need print? Let the library go the way of the bookstore. Disintermediation is inevitable.  Perhaps the answer lies in the library's physicality. Can the library do something a Kindle cannot?" 

Alternative Digital Recipe Libraries

As a long time cook this is a challenge.  Some alternatives.  But no, not impossible.  As simple as a file folder. Or as complex as an associative memory.

Text Analytics for Legacy BI

Always interesting Doug Lautzenheiser discusses in his business intelligence blog the use of text analytics techniques to search code libraries in enterprises for modernizing their function.  A novel approach that can greatly simplify this process to make sure that mission critical software is still doing things right.  I can think of many other kinds of searches that might be done in legacy 'knowledge' that could be useful to the enterprise.

Kindle Books Available from Libraries

Amazon has launched a Kindle book borrowing service from over 11,000 local libraries. You don't need a physical Kindle, only a free App for your device and an Amazon account.  Great news, I will experiment with the experience.    Press release here.

E-Books and Libraries

Good overview in ReadwriteWeb:  More indications of if libraries will survive and what form they will take.  Will they be in the Cloud with the rest of us?  Or parceled out in book like pieces? How will compensation be delivered to authors and publishers?   Will reading devices be checked out as well?  Seeing that here locally.How many e-copies will a library hold?  Lots of interesting questions to consider.  " ... There have been fears on the part of many librarians that the move to digital content would harm libraries - not because of any problem with the digital format per se, but rather that some of the restrictions on their access to materials and to lending would be too onerous. However it appears as though there are a number of new pricing models for libraries and a number of lending alternatives. Patrons clearly want to borrow and read e-books, and most libraries are in the process of developing programs to make that possible ... "

Amazons Lending Library

Librarians react to Amazon's lending library. " ... The stakes are incredibly high for public libraries right now. Federal, state, and local budgets are tight. Libraries are closing or cutting back on services. Alongside these fiscal trends are digital trends: the explosive growth in e-books, something that is radically changing the face of book publishing, book distribution, and yes, book lending.... "

Library Augmented Reality for Grocery Shelves

Found this Readwriteweb article on book shelf placement to be interesting.  I have worked with libraries on catalog systems.  It is not only libraries where things can become become disarranged. Shelves in grocery often do not conform to the prescripted plan-o-gram.  Is there a reapplication of this mobile library App to the complexity of a grocery shelf?  I can think of further information that could be leveraged at the shelf beyond position of an item.

" ... If you've ever worked in a library, you're familiar with the drudgery of shelf reading. That's the process of verifying that all the books on a shelf are in the right order, based on their call numbers. Books get out of order fairly easily, when they're taken off the shelf and examined, for example, or when they're just stuck in the wrong place.

Miami University's Augmented Reality Research Group (MU ARRG) ...  led by Professor Bo Brinkman, has developed an Android app that could save librarians a lot of time and hassle. Using the Android's camera, the app "reads" a bookshelf, and with an AR overlay, quickly flags those books that are misplaced. It will also point to the correct place on the bookshelf so the book can easily be re-shelved correctly ... "