Saturday, December 05, 2015

Presentation at Italian Drupal Day

I'm just back from Bologna, where I attended with my colleagues the Italian Drupal Day conference. We at Net7 are working on several Drupal based projects (the latest I've managed being the website of Scuola Sant'Anna, one of the most prestigious universities in Italy).

We decided to do a presentation on the project, in which we exploited the semantic API of Dandelion to completely automatise the work of an editorial team. Software services fetch articles from more than 40 web sites (in Italian, English and French) and analyse their texts using Dandelion's Named Entity Extraction and Automatic Classification services. If the article matches with the topics of interest of the portal, it is automatically published, if not it gets discarded.

The site has been in production for several months now, publishing hundreds and hundreds of selected contents, in three languages, without a hiccup and without any manual intervention.

The Drupal Day slides follow (in Italian). Enjoy!

Wednesday, August 05, 2015

My Nerdie Bookshelf - "Linked Data - Structured data on the web" by David Wood, Marsha Zaidman, Luke Ruth and Michael Hausenblas

This book has been a bit of a disappointment to me, the first one I had from Manning Publications.

Despite being published in 2014 you have the impression that the information provided here are stale. Only in the final chapter ("The evolving Web") a comprehensive, well written and updated viewpoint on Linked Data (and the Semantic Web) is provided, although in concise form.

The foreword by Tim Berners-Lee and the collaboration with Michael Hausenblas lured me to blind purchase the book. In particular I was looking for insights in what, in my viewpoint, is a powerful use case for Linked Data which hasn't been addressed enough, that is Semantic Enterprise Data Integration, hoping to get, as it is common for Manning books, a lot of advanced technical information. In particular I was, and still am, looking for technical advice, integration patterns and product reviews that can guide me in using Semantics to effectively interconnect enterprise data silos.

The book on the other hand revolves around a different perspective, those of a data publisher, with little (if any) notion of the technology behind Linked Data. It presents therefore all the basic concepts at a quite simple level.
This is of course a legit editorial choice but what annoyed me the most was the fact that the information provided are often outdated. No mention on JSON-LD or to the Linked Data Platform principles; CKAN, a widely used platform for creating open data repositories, is just cited but only in connection to the DataHub site. Moreover, the motivations, advantages, pros and cons of working on Linked Data are presented in a very basic, if not superficial, way.

The mention of Callimachus, the "Linked Data application server" created by the authors, left me unimpressed as well, even if it is correct to say that it has been used in interesting projects.

I must admit that I am biased and might sound arrogant (sorry if this is the case): at the end of the day I've been working on these topics for 5 years. The fact is that this book could have been appealing to beginners if only could present more up-to-date information and more detailed use cases. Linked Data looks like it was written in 2010: it could make sense to publish it in 2011, not, as it was the case, in 2014.

Friday, June 05, 2015

Introducing Social Proxy

I've finally published on SlideShare the presentation of Social Proxy, a project I've been working on since 2010.

If you ask, "why this platform and not HootSuite or Radian6?", well I think it still has some strengths, despite our (huge!) competitors have received tons of VC funds over the years, while basically Social Proxy has been developed through a series of orders (some very small) from our customers. In fact:

1. Social Proxy offers, in a single SaaS offering, plenty of features that you can only get by acquiring multiple services. You can get Social Media management (à la Hootsuite) and Social Media Analysis (see Radian6). It is certainly less advanced respect these famous competitors but... it still performs more than nicely!

2. Social Proxy is a framework for Net7, that can be easily extended when a new, custom feature is needed by a customer. For example this Drupal web site, doesn't have an editorial team behind. It presents content automatically fetched and "cleaned" through the Social Proxy: dozens of RSS feeds are scanned, the linked pages retrieved and their content is extracted, keeping the main text and removing all the decoration parts. Through web services, the Drupal site fetches and publishes the curated content.
Of course competitors provide APIs but the amount of things that you can do with them is limited.

Anyway, here are the slides: enjoy the reading! Other information on Social Proxy (in Italian) can be read here.

Monday, April 27, 2015

There’s Semantics in this Web

I was asked by Dr. Serena Pezzini of the CTL department of the Scuola Normale Superiore of Pisa to do a presentation on the Semantic Web on April 16th (beside there’s a photo of me taken at the event). Slides, in Italian, are available on SlideShare: the preparation has been a quite interesting process, so I thought to share it in this blog here, this time in English.

This presentation for me was in fact like opening up the legendary Pandora’s Box. It ignited a reflection about what we as a company do regarding the Semantic Web. Net7 in fact always characterizes itself as a “Semantic Web Company".

At about the same time I was contacted both by CTL and by a partner company to talk about this subject. On the one hand CTL expected suggestions and stimuli to use Semantic Web technologies in their work on the Digital Humanities field. The partner company was looking for professional training on these topics.

For 5 seconds I went into autopilot mode and started to think about explaining the Semantic Web in the standard fashion (RDF, ontologies, triple stores, SPARQL, RDFS, OWL, well, you got the idea…). Then three questions sprang to my mind...

Do these persons really need this kind of information? Are they going to really use all of this on their daily job?

The second question is a bit discomforting: do we at Net7 really use completely and especially consciously the whole of the Semantic Web technologies?

The third is even more serious: what’s the current state of the art of the Semantic Web? Is it still an important technology, with practical uses even for middle/low-sized projects, or should it stay confined in the Empyrean of research and huge knowledge management initiatives?

So, it was really important for me to do a presentation with the attempt to find answers to these questions, to present topics that could be of interest and useful to the audience and at the same time to put in a new perspective my knowledge on the field.

The presentation therefore came out as a reflection on the possible uses and advantages of "Semantics in the web", first and foremost for me, in order to reorder my mind, with the hope that it can be useful for others as well. I tried therefore to take a step back, hopefully to progress further in perspective.

For its preparation I read a great deal of material (see bibliography at the end) and was heavily influenced by the presentations and articles of Jim Hendler (not to mention the fantastic book “Semantic Web for the working ontologist” that he co-authored). So, even if you won’t read these lines, thank you very much Dr. Hendler for your insightful thoughts!

Coming back to my presentation, it is not a case that I used the concept “Semantics in the Web” and not “Semantic Web” in the title. Semantics in fact, in the light of all the readings that I did, seems to me more important than the technology behind it.

I started the presentation with a small historical digression, from the very first vision of the World Wide Web in the Tim Berners-Lee's 1989 original proposal, up to the seminal 2001 article on Scientific American, where Berners-Lee, James Hendler and Ora Lassila presented the Semantic Web.

I continued by explaining the key concepts of the Semantic Web, which served to prove how Semantics, despite the Semantic Web vocal critics, can still count huge success stories in the web of today.

The funny thing is that the Semantic Web’s vision didn’t exactly materialize as expected by its inventors. On the one hand is fundamental to comprehend how things in web history just happens through serendipity. On the other is crucial to have always in mind the Jim Hendler’s motto “a little semantics goes a long way”. Indeed just a small portion of the Semantic Web “pyramid” (see slide number 42 in my presentation, taken from a Jim Hendler’s keynote) finds a recurring use, while the rest (inferences and the most sophisticated OWL constructs included) has still a limited diffusion or is just relegated in high-end research initiatives.

So the Semantic Web hasn’t failed but materialized a bit differently than expected. One therefore should really think to Semantics first, that is to exploit the knowledge that can be extracted from documents, linked data repositories, machine readable annotations in web pages (SEO metadata included) before worrying about the orthodox application of the complete stack of Semantic Web technologies.

The Semantic Web is on the other hand a still promising and on certain aspects undiscovered territory. While I don’t honestly see it as a key technology to power web portals (there are plenty of more mature technologies, even open source, - think of Drupal or Django - that fit better this purpose) the idea of managing information through graph makes a lot of sense in several areas, including:
  • knowledge management with highly interconnected data (think of Social Network relationships). Here the capacity of triple stores to handle big graph data will really make the difference, especially if an open source product can be used for this purpose (recently we @ Net7 have bet on Blazegraph and while we have been quite satisfied until now, it must be also said that our graphs are not exactly “that big”). There is no doubt in fact that solid open source products are fundamental to skyrocket the use of specific technologies and software architectures (think of LAMP).
  • extraction of structured data from text: a great classic Semantic Web use case indeed
  • linking independent repositories of information, implemented with traditional technologies in multiple legacy systems (another Semantic Web classic).
  • raw data management and dissemination.
The latter is something that we @ Net7 would really like to explore in great details in the near future. The idea arises from a contact that we have with a local medical research centre: we noticed that very often their management of data acquired through sensors is untidy. This leads to data loss and corruption. Moreover, raw data after the medical research is over, gets dismissed. On the other hand if this data could be:
  • formally described in great detail
  • openly distributed, after a specific anonymizing process in order to remove “sensible information” from it
it might open the door for its reuse. This way scientists from all over the world can take this data and exploit it in their research, increasing the dimension of their data sets and consequently improving the probability ratio of their experiments. This isn’t something new indeed but it will become more and more relevant in the near future since the European Commission is fostering Open Access to research data in the Horizon 2020 projects.

I concluded my slides by also noticing that Semantics is becoming more and more a commodity, offered through specialized cloud services. Named Entity Recognition SaaS offerings, SpazioDati’s DataTXT and AlchemyAPI included, are a consolidated reality. Cloud Machine Learning services are becoming mainstream (see in this regard this insightful article on ZDNet). Developers therefore can enjoy “a little semantics” in their application, without embracing the Semantic Web in full. As Jim Hendler says in fact… a little semantics goes a long way!


Tuesday, February 17, 2015

“The importance of being semantic”: Annotations vs Semantic Annotations

The focal point of many Net7’s projects is on semantic annotations. What does it really mean and why the term semantic is so important in this context?

Annotations can be simply seen as “attaching data to some other piece of data”, eg documents; the advantage of a semantic annotation is to have this data formally defined and machine-understandable. This offers better possibility for search, reuse and exploitations of the annotations performed by users.

In [Oren 2006] a formal definition of annotation has been proposed. Simply speaking it consists of a tuple with:
  • A. the information that is annotated inside the text
  • B. the annotation itself
  • C. a predicate, that establishes a relationship amongst the two points above
  • D. the context in which the annotation has been made (who made it, when, its reliability, a possible limit for its validity, etc).
This holds true for all kinds of annotation, including handwritten notes beside paper documents.

A simple definition of semantic annotation is proposed in [Kiryakov 2004], where it is presented a vision by which “...named entities constitute an important part of the semantics of the document they are mentioned in… In a nutshell, semantic annotation is about assigning to the entities in text links to their semantic descriptions”.

A very effective definition of semantic annotation can be found in the Ontotext web site: “Annotation, or tagging, is about attaching names, attributes, comments, descriptions, etc. to a document or to a selected part in a text. It provides additional information (metadata) about an existing piece of data. … Semantic Annotation helps to bridge the ambiguity of the natural language when expressing notions and their computational representation in a formal language. By telling a computer how data items are related and how these relations can be evaluated automatically, it becomes possible to process complex filter and search operations.

This is of course a step forward respect traditional Information Retrieval techniques, in which documents are managed (and indexed) as a disarranged “bag of words”, with no attention to their meaning and no ability to identify ambiguities due to synonymy or polysemy.

Ontologies (or more simply speaking “vocabularies”) provide a formalization of a knowledge domain. Some of them are generic (like for example OpenCyc or and can be used to provide meaningful, albeit not domain-specific, descriptions of common facts and events.

A simple example of a "movie" ontology, taken from the Internet, is depicted below: it presents entity types/classes (Movie, Person, Character), their attributes/metadata (title, name, gender) and relationships among entities and attributes (HasCast, Is directedBy, etc). Of course for a professional use is hugely important to provide very specific vocabularies, that describe in great detail a certain domain.
Movie Ontology Example - source:

The key for providing effective “semantic descriptions” through annotations therefore lies in:
  • the careful definition of ontologies, that is the vocabulary of terms, classes, predicates and properties according to which semantic annotations are performed by users;
  • the use, as much as possible, of standard ontologies, whose meaning is therefore well-known and accepted. This allows the automatic interpretation of the metadata defined using them;
  • the exploitation of Linked Data in annotations. It’s quite convenient to use standard, well-known and formally defined web datasets in the annotation process. Datasets as Wikipedia (or better, DBPedia, its Semantic Web version) or Freebase provide both huge, general purpose, vocabularies of entities and terms, that can be referred through “linking” in the annotation process, and a sets of well-known ontologies, that are so common to be considered standard and as such easily understandable by semantic-aware software agents.
For all what has been said so far, there are several advantages that semantic annotations provide, for example:
  • they are machine understandable since, reusing the definition of [Oren 2006] presented above:
    • the predicate C is formally defined in an ontology
    • the type of the annotation B is formally defined in an ontology
    • the annotation B itself can be an entity formally defined (in an ontology or in a public dataset, eg Wikipedia)
    • the context D may be formally described with terms, types and entities from ontologies and public datasets.
  • they precisely define the context of the annotated document, identifying in detail the nature of the information that is under study. This can be exploited in searches, classification and more generally in every possible reuse of annotated data.
  • they open the door to inferences, that is deducing automatically other data that relate to the annotated document and complement/enrich the annotations originally performed by a user.
The importance of semantic annotations in text can be better appreciated through an example. Consider the following three sentences:
  • On the morning of September 11 Australian swimming legend Ian Thorpe was on his way to the World Trade Centre for a business meeting.
  • Born on September 11 writer/director Brian De Palma’s career began to take off in the 1970s with the horror classic Carrie, based on a Stephen King novel.
  • On September 11, President Salvador Allende of Chile was deposed in a violent coup led by General Augusto Pinochet.
Even if they all contain a reference to the same day (September 11), its actual semantic in them is very different: the first one is an Event (in 2001), the second a Birth date (1940) and finally the last one refers both to a Death date (Allende’s) and to an Event (Pinochet’s Coup in Chile in 1973).

A simple textual indexing of these three texts would identify the “September 11” fragments, without any understanding of their meanings and the actual years they refer to. On the contrary, by simply annotating them with a link to the corresponding entities of Wikipedia (namely:
September 11 attacks, Brian De Palma, 1973 Chilean coup d'état) and the use of specific predicates (Eg. dbpedia-owl:eventDate, dbpedia-owl:birthDate, dbpedia-owl:deathDate), one can automatically infer the correct contexts plus much more data.

Finally annotations can be performed manually by users or automatically, using software services that can identify terms in a text and associate them, through a specific predicate, to entities of a controlled dictionary or of a linked dataset.

Why all this talking about Semantic Annotations? Well, the reason has a name: Pundit!

Pundit is a web tool that allows users to creare semantic annotations on web pages and fragments of text, with an easy to use interface. Pundit is the foundation stone of the PunditBrain service in the StoM project and of many other Net7 initiatives, both research oriented and commercial.

Although Pundit is mainly used for manual annotations, it already supports automatic entity recognition by using several software services, including DataTXT, a commercial service of SpazioDati whose main development has been carried out in the SenTaClAus research project in which Net7 was also involved.

The 2.0 version of Pundit, currently (February 2015) under development and in "alpha", can be tested on the project web site. Hopefully soon a demo version of PunditBrain will be also released to the public.