Pandemic Atlanta Lockdown - Week 16

This was a long weekend week. I almost forgot it! Anyway, when a long weekend comes, we still try to compress the whole week of meetings and work into 4 days, and as such the difference becomes less.

My personal bar currently has:
Everclear 95%
Cherry Moonshine 50%
ABSOLUTvodka_US 40%
Malibu black Coconut 35%
Crème de Menthe 30%
Crème de Cacao 24%

@Kahlua 20%
Amaretto 20%
Irish cream 17%
Peach Schnapps 15% 

and 

Grenadine Syrup
Org orange juice
Heavy cream
Milk
Vanilla ice cream.

 

I also got the notification that OTM/CoopIS 2020 has been canceled. I was planning to submit a paper there, after our successful presentations there for CoopIS 2015 and 2016. This is quite sad. This is the first conference canceled due to COVID-19 that I encountered so far. All the other conferences managed to go virtual.

Unfortunately, we have to cancel OTM and hence CoopIS 2020

Dear Pradeeban Kathiravelu,

As PC chairs for the CoopIS conference we regret to let you know that the conference will not run finally during this year.

Taking it from the whole set of conference organisers: "As you might appreciate some of the people on whose effort we must rely for organising a complex quality event just could not free the time for this as they had to cope with, in some cases, absolute family survival. Also voluntary work understandably is in very short supply these days."

It seems to have been a hard decision from the organising committee, where a set of circumstances (including family-related ones) have provoked the final decision, and we are sorry that you had already invested your time on preparing a paper for it.

We hope to be able to resume next year.


On July 13th, I received another email from easychair:

Cancellation of CoopIS conference edition in 2020

OTM 2020

Sorry if you have already receive a previous communication from us. This communication is meant to be for those authors that have submitted their contributions in the last week.

Unfortunately, the organising committee of the OTM series of conferences has decided to cancel the conference series, something that is beyond the control of us as PC chairs. We have tried to communicate it in as many places as possible (including the website, which is currently down), and we are sorry that you have been working on your submissions these days. We have not been able to cancel the easychair installation either.

Regards, and apologies for the inconveniences,

  

So it seems there were some unfortunate creative differences between the conference organization due to COVID19. Hope things will get better soon. 

Remote conferences - Conferences during the COVID19 Era

Lisboa, 2013 NY, and my Zoom background

COVID19 has made the conferences go remote. Some conferences, such as SDS 2020, were postponed in the hope that COVID19 will end soon. But eventually, COVID19 became a deadly pandemic that no conference could survive. Some conferences, such as DisCoTec 2020 became free to attend for any participants while charging the authors a negligible fee. Other conferences offered discounted rates, such as 50%. For those who have already paid the full registration fees, refunds for the overpayment were issued.

DisCoTec used the Zoom webinar for the conference. Unlike a Zoom meeting that lets you see and listen to every participant, Zoom webinar gives a monologue-kind of environment. Most, if not all, presenters had recorded their presentation as videos. It is similar to the SDS 2020 presentation that I am supposed to give by the end of June. We all have recorded PowerPoint presentations with our voice over and uploaded them as videos for the conference. DisCoTec had a similar approach, with the videos uploaded to YouTube, unlisted. I mostly attended the tutorials (Monday), and DAIS sessions (Tuesday - Thursday). I skipped the workshops (Friday).

There are several advantages in conferences going remote. First, remote online conferences let us continue conferences as usual, without giving up. Now the options are: remote conferences or postpone or cancel the conference. With the pandemic still going strong, postponing does not seem to work. That leaves us with either no conference for 2020 or a remote conference. Something is better than nothing! So remote conferences, it is. Second, they open up the conferences to a wider community - those who couldn't travel to conferences due to funding and visa limitations. Now, with less funding, we could join more conferences. OK, these advantages are quite obvious.

Now, let's look into the shortcomings of a remote conference, done in the DisCoTec way.

1. No scientific interaction between participants

Following the recorded presentations, there were live Q&A sessions that lasted up to 5 minutes. But, mostly, it was just listening to recordings. With time, I am sure the conference will list those uploads, making it possible for everyone to go and watch them on YouTube, on-demand. To incentivize attending to the conference session remotely, conferences must put additional efforts, beyond the Q&A. Otherwise, I may just watch them at leisure, and send my questions as emails to the authors.

Of course, the conference organizers created a slack channel (on Tuesday) for questions and answers from the sessions. Each conference session had its room. DisCoTec is an umbrella event with a parallel conference (including DAIS) and workshop sessions. So it had several rooms in the Slack channel with one general channel. 

2. No coffee-hour friendships and networking

One thing I usually love about conferences is the coffee hour discussions. We meet the presenter and ask questions or compliment their work. I am an extrovert when it comes to conferences. I roam around with a cup of coffee during the coffee breaks. I go and introduce myself to any lone participant and discuss their work. Penetrating an already existing group is harder. Sometimes, participants from the same university, company, or friends from the same country make a group and stick to each other. I always avoid that. Such groups usually appear less open to make new friends. On the other hand, I have always made friends with diverse groups that were formed on the spot. I have made groups at conferences, by joining several participants over the coffee hours. Then we all go for dinner or sightseeing together. The best of the conferences, indeed!

3. Less opportunity to find long-term collaborators

Conferences are an excellent opportunity to find long-term collaborators and even employers! During IC2E/SDS conference, I met my supervisor from Croatia for the first time. She had attended my presentation and found that her research aligned with my on-going Ph.D. research (which was still in its infancy). That's how I ended up in Croatia, spending a summer in a short-term scientific mission (STSM)! That STSM contributed to the core of my Ph.D. research. Such fruitful interactions won't happen easily in online conferences. At least not in this format. I tried to interact with participants on Twitter via the hashtags. But not much luck except for some minor chats. The potential to meet collaborators is highly limited in remote conferences such as this and almost non-existent for the audience. I hope that at least presenters got some visibility and interaction beyond the Q&A. Communications with potential future collaborators are the key motivation for conferences. Otherwise, I would rather publish my work in a journal instead. In fact, journals are even valued highly in academia. So there must be more incentives to submit and present our papers in conferences when the advantage of meeting researchers physically is no more.

4. Timezone differences

The timezone difference is the biggest challenge, even for an introvert who doesn't like to network with random researchers. The conference was in Malta timezone, as it was originally scheduled to be in Malta. So for the 10:00 - 18:30 sessions, I had to wake up early in the morning, as it is 4:00 - 12:30 in Atlanta time. I was planning to sleep early and wake up early. But I also had other work meetings, and I could not escape my work duties for the week. Hence, I ended up having to spend most of the week without proper sleep. I went to bed at regular times and woke up as early as 4 a.m. Given a chance, I would go to bed earlier and adjust my timing to the conference timezone, including the meal hours. That would have made me virtually travel to Malta!

Although when we travel to conferences, I still end up with little sleep - as I have to prepare my presentations and try to find some time to explore the city. I am used to timezone shifts. But the last week was like double work since I had the conference sessions followed by my regular work hours. If the conference sessions were held at Atlanta time, of course, I would have to sacrifice one of the two - either the work for the week or the conference.

Can we do better?

Yes. My university in Portugal (IST/Instituto Superior Técnico, Universidade de Lisboa) is hosting remote seminars on Mathematics, Physics & Machine Learning weekly. I joined the session on "Learning from Distributed Datasets: An introduction with two examples" last week. It was interactive. The session was hosted in a Zoom meeting. So the participants were able to see each other and even interact with the presenters before the event. It was nice to introduce myself to my ex-colleagues remotely. It was my first IST seminar after two years. Also, it is my first remote IST seminar ever. These online open seminars are one good thing that happened out of the pandemic. The recordings are, of course, available to watch later.

So in my opinion, such Zoom meetings that allow the participants to see and interact with each other before the sessions and during the "virtual coffee breaks" is a better option. Of course, zoombombing will be an issue. Password-protected conference zoom meetings can fix that.

The current move towards online conferences to combat COVID-19 travel limitations is a welcome move. However, it is not sure how long this will last. It will depend on the COVID-19 as well as the associated travel restrictions. The conference committees should take extra measures to make the conferences more interactive, to encourage communications between the participants and networking, rather than merely playing the recorded videos followed by a customary Q&A. This is new to all of us. So, of course, this is going to take a lot of iterations. Moreover, I empathize with all the conference organizers. They all have made tremendous efforts to organize the conferences physically. But they had to cancel their plans and again prepare for the online conferences.

Now CoopIS 2020 has gone online. It is one of the conferences I have presented in the past (I had papers there in 2015 and 2016). I am looking forward to attending more remote conferences whenever the topic is relevant to me, as long as the registration is free for the participants. I am yet not sure whether paying the registration fee is justified for a remote session, unless my paper is published in the conference.

Selective Redundancy in Network-as-a-Service: Differentiated QoS in Multi-Tenant Clouds

I arrived at the beautiful Greek island of Greece this Sunday to present one of my papers at CoopIS conference and another at EI2N workshop. Today I presented my first paper at EI2N. The presentation and abstract are given below.

Selective Redundancy in Network-as-a-Service: Differentiated QoS in Multi-Tenant Clouds from Pradeeban Kathiravelu

Abstract: Data centers consist of various users with multiple roles and differentiated levels of access. Tenant execution flows can be of different priorities based on the role of the tenant and the nature of the process. Traditionally enterprise network optimizations are made at each specific layer, from the physical layer to the application layer. However, a cross-layer optimization of cloud networks would utilize the data available to each of the layers in a more efficient manner.

This paper proposes an approach and architecture for differentiated quality of service (QoS). By employing a selective redundancy in a controlled manner, end-to-end delivery is guaranteed for priority tenant application flows despite congestion. The architecture, in a higher level, focuses on exploiting the global knowledge of the underlying network readily available to the Software-Defined Networking (SDN) controller to cater the requirements of the tenant applications. QoS is guaranteed to the critical tenant flows in multi-tenant clouds by cross-layer enhancements across the network and application layers.

Data Quality and ∂u∂u

Data Quality is a course I did at the university, as part of the 5 courses requirement for my PhD.  This course was remarkable at so many levels. 

First, I had to go to Taguspark after two years, for this course. Second, I really learned a lot. Third, I was the only student for the course. Usually individual courses are expensive. Perks of being a PhD student. Finally, we had a full paper at CoopIS 2015 (CoopIS is an "A" rated conference), as a result of this course work, which was awesome. I scored 19 out of 20 for this course. I read a lot of papers for this subject, and I spent majority of my time for this module in my second semester. Good memories. :)

Read more about our paper here. The full paper can be accessed here.

∂u∂u Multi-Tenanted Framework: Distributed Near Duplicate Detection for Big Data

Our paper "∂u∂u Multi-Tenanted Framework: Distributed Near Duplicate Detection for Big Data" Pradeeban Kathiravelu, Helena Galhardas, Luís Veiga was presented at the 23rd International Conference on COOPERATIVE INFORMATION SYSTEMS (CoopIS 2015), 28-30 October 2015, Rhodes, Greece.

Due a clash in my travel schedules, I could not attend the conference and present the paper myself. Hence, my supervisor and co-author Prof. Luís Veiga presented the paper in Greece on the 29th of October.

Abstract.
Near duplicate detection algorithms have been proposed and implemented in order to detect and eliminate duplicate entries from massive datasets. Due to the differences in data representation (such as measurement units) across different data sources, potential duplicates may not be textually identical, even though they refer to the same real-world entity. As data warehouses typically contain data coming from several heterogeneous data sources, detecting near duplicates in a data warehouse requires a considerable memory and processing power.

Traditionally, near duplicate detection algorithms are sequential and operate on a single computer. While parallel and distributed frameworks have recently been exploited in scaling the existing algorithms to operate over larger datasets, they are often focused on distributing a few chosen algorithms using frameworks such as MapReduce. A common distribution strategy and framework to parallelize the execution of the existing similarity join algorithms is still lacking.

In-Memory Data Grids (IMDG) offer a distributed storage and execution, giving the illusion of a single large computer over multiple computing nodes in a cluster. This paper presents the research, design, and implementation of ∂u∂u, a distributed near duplicate detection framework, with preliminary evaluations measuring its performance and achieved speed up. ∂u∂u leverages the distributed shared memory and execution model provided by IMDG to execute existing near duplicate detection algorithms in a parallel and multi-tenanted environment. As a unified near duplicate detection framework for big data, ∂u∂u efficiently distributes the algorithms over utility computers in research labs and private clouds and grids.

The full paper can be accessed here.

∂u∂u Multi-Tenanted Framework: Distributed Near Duplicate Detection for Big Data from Kathiravelu Pradeeban