Language Technologies and Learning Sciences

There is an interesting area emerging at the cross section of learning sciences and language technologies. Applying various language technologies including intelligent information retrieval, information extraction, personalization, summarization, and cross language information access to create or enhance an effective learning environment seem to be of immense benefit and promise.

Some of our (SIEL’s) past and current work can be applied to explore the role of personalized search and personalized summarization technologies in creating collaborative leaning environments.

We plan to collaborate with LTI of CMU in this area. Our project on Creation of Smart space in the classroom (funded by Nokia University Relations Program) also falls in this area.

A Paper Into ACL 2009

Here gets a paper ACCEPTED into acl-ijcnlp-2009.

Title:

“QUERY-FOCUSED SUMMARIES OR QUERY-BIASED SUMMARIES?”
By: Rahul Katragadda

Hearty Congrats From SIEL, LTRC, IIIT-H  to Rahul Katragadda.

Checktronix India Pvt Ltd.

Here i am writing my experience with the Checktronix Interview Process. 

About  the company – look here

The Process is as follows:

1. Technical written – Domain (NLP , Machine Learning, IR & IE) , Data Structures, Programming.

2. Psychometry Test/ Attitude Test – here they show few slides of pictures, asked to write what their mind gets strike once see the picture and at the end they will ask to write story with the picture. This test is to check the attitude.

3. Interview process is different for R&D and Engineering side.

For R&D, they mostly look for domain knowledge in NLP. If you have the better knowledge in the following topics then your interview wil be very easy

1. Supervised, Semi supervised, unsupervised learning

2. Inf.Ext, Inf.Retr

3. HMM, CRF

4. LSI

5. LDA, PLSA, CTM

6. Bayesian

7. Perl, C++

8. Data Structures

9. KNN, K-means, Fuzzy-C, Clustering algos

For Engineering side, they look for more logical thinking and they ask more from the academics. In first written exam there will be C,C++,Java questions, no domain questions.

Any queries , let me know.

Note: Share your interview experience and the questions asked. This will be helpful for everyone.

Hysea Cloud Computing Workshop

Yesterday’s HYSEA event on Cloud Computing (CC) at Google office was very successful. I have presented a few academic perspectives on the usage of cloud computing as well as some research challenges in this area.

For me, there are a few interesting takeaways from this workshop:

• The importance of this area is clear from the very fact that Amazon, Google, Yahoo, Microsoft, IBM, HP, Sun investing heavily into this and the pace at which tools, products, services and applications are being built.
• It appears that the landscape is being filled with various players focusing on their own specific sub-area. For example, Google App Engine serving the developer/programmer community, Amazon EC2 and S3 serving the application development infrastructure, and Microsoft Live Mesh focusing on consumer solutions.
• The need to create new progamming paradigms in cloud computing area is apparent, which includes creation of libraries to adopt existing programming languages to suite the CC style programming and creation of new languages for service descriptions and for monitoring (at the Service Level Agreements and beyond).

I have used the term “Cloud Sourcing” to describe the trend of delegating the software development, production and maintenance beyond the boundaries of an enterprise. It is clear that the CC playground is wide open and it will be interesting to see who will emerge as a winner. Like India emerged as a superpower in outsourcing, someone (country/region/company/…) will become a super power in Cloud Sourcing.

BTW, Hyderabad is buzzing with cloud computing activities – including a barcamp on August 30th at Amazon.com Development Center. Check it out if you are interested.

Google Tests Related Search Phrases Inline With Search Results

It was noticed that Google is testing a new location for “related” search results. The location appears to be directly under each individual search result’s URL.

You can find the snapshots here.
http://www.seobook.com/google-beta-testing-showing-related-phrases-near-documents

Cloud Computing and IR

Here is a post on Search Engine Watch (SEW) today:

Cloud Computing Unites Yahoo, HP, Intel

Yahoo, Intel and Hewlett Packard announced an alliance to advance “cloud computing,” backing a global trend that threatens Microsoft’s iron-fisted grip on packaged software installed on computers.

Earlier this year Google and IBM teamed up to advance research into providing SAAS (software as a service) on the Internet hosted by data centers.

Cloud computing isn’t as futuristic as it sounds. Web-based email (Gmail, Yahoo Mail, Hotmail) offered by Google, Yahoo, Microsoft is an example of cloud computing. Google has expanded its online software offerings to include word documents, spread sheets and more under the Google Docs brand.

The cloud computing initiative falls in line with Yahoo’s goal of open systems. Results of research done at the centers will be made public to make it easier for software developers to write applications for cloud computing.

“We believe this collaboration will do a great deal to take the research to the next level,” Yahoo Research chief Prabhakar Raghavan said during the conference call.

“We are really fueling the ecosystem here… Inevitably an application developer will build more readily for the cloud.”

—–

Guys, reserve your seats in the course on “Topics in Software Architecture” CS6600 this semester. The theme of this course changes every year and for this year it is “Cloud Computing”.

There will be guest lectures from Amazon, Microsoft, and Yahoo. BTW, There are only 20 seats available

Cuil Launches Biggest Search Engine on the Web

Cuil, a technology company pioneering a new approach to search, launches its innovative search offering, which combines the biggest Web index with content-based relevance methods, results organized by ideas, and complete user privacy.

Summary of Cuil’s features:

• Biggest Internet search engine—Cuil has indexed 120 billion Web pages, 3x more than any other search engine
• Organized results—Cuil’s magazine-style layout separates results by subject and allows further search by concept or category
• Different results—Unlike other search engines, Cuil ranks results by the content on each page, not its popularity
• Complete privacy protection—Cuil does not keep any personally identifiable information on users or their search histories

For more information, refer the following links.

http://www.cuil.com
http://searchengineland.com/080728-000100.php

Summer Activities in NLP/IR

As the summer break is coming to an end, all the SIELers are geared up to start the fresh semester with a sense of accomplishment, I believe. This summer, many of us are busy with various activities such as preparing for TAC and CLEF competitions, participating in summer schools and working on our projects (which is evident from the lack of posts in this blog)

We have submitted three runs for update summary task of TAC and also one for QA task. Let us keep our fingers crossed.

There were three summer events in NLP/IR areas where I gave a few talks:

• IASNLP at IIIT-H,
• NLP-Summer school at IIT Kharagpur and
• NLP-Ontology Summer workshop in IIT Bombay.

Each of these workshops were wonderful and I really enjoyed interacting with the participants. I focused on  IR/IE at IIIT-H workshop with a few examples on building scalable IE systems. My talk at IIT Kharagpur built on previous talk with a focus on recent trends in IE including the trend towards building frameworks. In IIT Bombay, I talked about personalization in IR/IE and IA.

Inner Game of Innovation – 2nd Part

Most people think that disruptive innovation is the only significant type of innovation and all other types of innovation cannot be treated at par with this one as they are somewhat inferior. I would have agreed with this view had I not come across “Dealing with Darwin” (by Geoffrey Moore), the book I talked about in the previous blog with the same title.

When I learnt that there are several types of innovations categorized into various zones in the life cycle of a product or a technology. These zones include: Product leadership (birth and early struggle of the product/technology), Customer Intimacy (where the product is making into the heart of the customer, making its space in the customer mind and life), Operation excellence (build the same product better, faster and cheaper - here is the zone where Indian software service companies showed a lot of innovation) and finally Category renewal (where you know when, where and how to stop the technology and move all the lessons learnt into something new which takes you to the next level). In these four zones there are about 15 types of innovations.

I have focused mostly on these aspects when I finally gave my talkon Innovation at CA (on 29th January). It was a great learning experience for me as I thoroughly enjoyed my preparation for this talk.

If some one is interested in learning more about these concepts they should buy and read this book AND refer to Geff’s website:

www.dealingwithdarwin.com 

IJCNLP-2008 Conference

Today is the beginning of NLP festival (all conferences are like festivals as we celebrate our research and have loads of fun) right here in IIIT Campus and I am happy to see most of SIELers taking very active part as organizers, volunteers and participants.

There is one paper accepted in the main conference (a matter of pride for all of us) – this is based on Rohini’s work on personalization (many of you might know that U Rohini after graduating from here joined AOL R&D, Bangalore) – this paper will be presented on 10th of January 2008, between 15:30 – 17:00 Hrs in session A3.3 at ISB. The paper is:

“Statistical Translation Models for Personalized Search” – Rohini U, Vamshi Ambati and Vasudeva Varma

If you are participating in the conference, please attend this presentation.

We are also organizing 2nd International workshop  on Cross Language Information Access (CLIA-2008) along with IJCNLP (W4). For more information see the website of the workshop:

http://search.iiit.ac.in/CLIA2008

See you all there - and we will be looking forward to hear your experiences.