I recently came to know that if one (on J1 visa) wishes to travel to a foreign country from US and if returning to the U.S., the residence permit / visa must be valid for at least three months exceeding the last day of the intended stay in foreign country (1,2).
This is not a small ask!
Many J1 visa holders have residence permit of one year. Let’s imagine, your research group is super efficient, you managed to get a result worth publishing in 6 months, you submit the work to a good conference and here goes another 1-2 months in review process and finally, your paper is accepted! The conference is in 2 months and you are super excited! This is the best case in research.
Oh wait! Let’s count the months, 6+~2+2 =10 months already gone by the time you will begin your travel. But now, you cannot travel to a foreign country and return back to US from there since you will only have 2 months of valid stay left on your J1 visa.
Now, one can choose to either forgo in-person visit or travel to your home country from the foreign country and get your visa renewed. So far so good! However, it can get worse if you discover that the passport is going to expire soon. Unlike B1/B2 visa, J1 visa is not issued beyond one’s passport validity. So even if you have DS 2019 for a period of 2 years, you will only get J1 visa upto the expiration date of your passport. Please note, this is the case for India. There is an additional 6 month passport validity requirement for some countries (3).
If you are on this page, I need not narrate the scary stories of visa appointments. Ideally, one should renew the passport and get the complete extension at one go if possible. Renewing passport (tatkal) in India demands appointment booking at least a month in advance (in Delhi and Blr) (4). As J1 visa application can be started only after getting the new passport, one can’t book the US visa appointment in advance. This option is rather troubleseome and has two unknown variables i.e. Passport Appointment slot and J1 Appointment slot. If you prefer to be in control or have limited time to spend in India, then you can try the option of Indian passport renewal in US.
VFS Global provides the service of passport renewal either by post or appointment. Their website explains everything clearly and there are only few points which I wanted to add here:
That’s it. You are all set to post your passport renewal application. You can track the application status on VFS website (6).
Disclaimer: The information provided on this website does not, and is not intended to, constitute legal advice; instead, all information, content, and materials available on this site are for general informational purposes only.
References
]]>bundle exec jekyll serve
, I faced this issue. I don’t have the exact error messages. So please decide for yourself.
The error message suggested to execute bundle install
. However, this spiralled to other errors related to unf_ext
, commonmarker
and other packages suggesting gem pristine
installs. On trying the suggested solutions, I eventually got a message about not having the right permissions. These are the links which I went through during my search:
Finally, this is what I did:
which ruby
and which gem
brew doctor
Your system is ready to brew
. If any warnings are shown, first try to resolve them. In my case, I uninstalled homebrew
completey using this LINK and reinstalled using this LINK..zshrc
and .zshprofile
files.brew doctor
and verify that it returns Your system is ready to brew
.brew install ruby@3.0
or the version you prefer. If you run brew install ruby
, then it installs ruby@3.2
and eventually this will cause tainted object
compatibility errors with jekyll theme.
echo 'export PATH="/opt/homebrew/opt/ruby@3.0/bin:$PATH"' >> ~/.zshrc
, export LDFLAGS="-L/opt/homebrew/opt/ruby@3.0/lib"
and export CPPFLAGS="-I/opt/homebrew/opt/ruby@3.0/include"
which ruby
and which gem
sassc
. Pay attention to this warning and accordingly add gem 'sassc-rails'
in the gemfile in local repogem install jekyll
gem install bundler
bundle update
and bundle
bundle exec jekyll serve
Hope it was helpful.
]]>I started applying in Jan 2022. After numerous applications and rejections, I finally received postdoctoral fellowship offers from University of Pennsylvania, The Alan Turing Institute and Aalto University. It’s been almost a month since I landed in Philadelphia for my postdoc at University of Pennsylvania. And, I am so glad to share that the journey has been full of learning and quite a few challenges. Through this post, I would like to share a few points which I found helpful during my application process.
The first big challenge is, being aware of professors, research groups which are of interest, and their ongoing work. Only when you have an idea about the kind of research groups you would like to join, you can start preparing your application. ACL conferences/workshops, ICWSM, EMNLP , IC2S2 are a few highly reputed venues and one may start with going through the work being published. There are numerous research groups doing wonderful work in CSS and NLProc. So, I suggest that you spend a good amount of time in identifying groups/researchers whose work inspires you. In addition,
Identify mailing groups: During my search, I found the “corpora@list.elra.info” emails most helpful. This list has the most proactive group of people who actively post opportunities (phd, postdoc, faculty positions) in NLProc and CSS. You must subscribe to this group if you are looking for academic opportunities in NLProc. Machine Learning News Google group is another such resource.
Check for Job boards and domain specific platforms: In addition, I actively looked for positions on NLP job boards such as NLP People and ACL Employment Opportunities along with regularly tracking open positions on the websites of research groups of my interest.
Stay Connected: Being on Twitter and following the right research groups/researchers of your interest also help a lot. The postdoc applications often have an early deadline and the sooner you know, the more time you will have to customize your application. Moreover, Twitter, LinkedIn, ResearchGate are a few platforms to connect, see what’s going on in your field and be aware of the trends. You can always start with “following / connecting” researchers whose work is interesting to you and then check who else they follow.
Communication is the key. Don’t hesitate to contact the research group / professor if you have any questions about the ongoing research, or even the expectations from the postdoc. In fact, I suggest you to not wait for the advertisement if you are sure about a research group. Just introduce yourself and enquire over an email or even DM them on Twitter. I had a very positive experience, professors were super warm and even up for a meeting sometimes.
Once you have identified the research groups / positions, the next step is to prepare your application. In the advertisement itself, the list of documents for application will be clearly stated. The documents are often:
For some positions, there will be a dedicated email id to send the documents whereas for others, you may be directed to a job portal where you will have to fill in the details and upload documents. If you are applying for postdoc positions in industry, it’s a good idea to let your references know in advance. The application is not considered complete without the references. When preparing CV and research statement,
It is definitely worth investing time and effort in customizing CV and research statements for every new position. There are a lot of sample research statements/CV online which can be referenced to understand what to write and what not. You may also ask your peers from similar domains to share their first impression on your research statement.
You can expect to hear soon about the status of your application after the deadline. It is sometimes the case that negative results are not conveyed. Until and unless it is explicitly stated in the advertisement, you can always enquire about the status of your application if you feel it’s been sufficiently long and there is no news on the research group website. Like I said, don’t shy away from communicating your concerns and questions.
Hiring process for postdoc position is often quick and does not give you much time to prepare once you are in the process. It is thus a great idea to
The interview could be with the head of the group or a team of members in the research group. I personally experienced both type of panels. I was given the choice to pick a slot of my convenience out of available slots. Due to possible different time zones, it is a good idea to give a thought when picking a slot. If possible, opt for day time, and keep yourself free before and after the interview to avoid feeling rushed. If not explicitly informed, it is always good to ask in advance if the panel will prefer a set of slides or if there is a programming exercise to complete. Please keep in mind the time duration when preparing slides. Being stopped in the middle due to shortage of time, may defeat the whole purpose of slides.
In my case, I was given a programming exercise before the interview for the position of “Research Associate” at The Alan Turing Institute. Read the description carefully and think of evaluating criteria behind the given task if not mentioned. Design your solution accordingly. For instance, in my case, accuracy was not the concern and the panel wanted an interpretable model. During the interview, I was asked to explain my approach through slides and code notebook.
During the hiring process at The Aalto University and University of Pennsylvania, I only presented my slides discussing my recent work with a focus on skills stated in the initial adverstisement. It is thus always good exercise to revisit the advertisement before interview to ensure that you are covering everything. Another point to note is that these interviews are a good opportunity to connect and make a sincere impression. Make sure to stay connected with the research group/ professor to stay updated. Even if you receive a negative outcome, you can still learn from them. Bless the Internet and social media.
In the end, do remember that the interview is also for you to judge the research group and how useful it will be in furthering your objective whether it’s finding a job or just research experience in new domain. It is also a good idea to check the career path of ex-postdocs from the research lab. So, do your research, prepare your questions and try to get an idea how the research group supports their postdocs whether they choose to stay in academia or switch to industry jobs.
]]>The motive behind this study was to understand how the pattern of needs has evolved from pre-COVID-19 phase to different stages of COVID-19 namely lockdown, the first wave and the second wave. Indian tweets posted between Dec’19 to Jul’21, spanning a period of 19 months were collected using snscrape.
Manual annotating tweets with their expressed need is expensive and time consuming. We use topic modeling to label tweets - topic modeling is statistical modeling for discovering the abstract “topics” that occur in a collection of documents [wiki]. Idea is to identify these abstract topics in large collection of tweets and then label topics which are much smaller in size! As a proof of concept, we used Maslow’s Hierarchy of Needs (MHoN) (Maslow and Lewis, 1987), which categorizes the human needs into five distinct levels namely physiological[L1], Safety [L2], Love and Belonging [L3], Esteem [L4] and Self-Actualization [L5]. The first two levels namely physiological and safety in MTM are often considered as basic needs whereas the next three levels namely love and belonging, esteem and self-actualization are regarded as advanced needs.
Here, we illustrate the week wise distribution of tweets tagged with needs. One such tweet annotated with physiological need is “Nobody staying at hotels, So why not convert them into covid centers”.
The next task was to identify the tweets that were discussing unfulfilled needs. In Frustration-Aggression theory, Dollard et al. (1939) defined frustration as an impediment or blockage in achieving one’s needs or goals. We hypothesized that unmet needs can be detected by identifying if tweets already marked with needs contain frustration or not. We finetuned the pretrained RoBERTa language model for this task.
Below are two example tweets:
HOW fast does one have to be to book a slot on COWIN? I saw slots available at a hospital; I selected the time slot; entered the CAPTCHA in not more than 15 seconds... and still it didn't book the slot. And then when I refreshed, all the slots were gone.
- Frustrated
I did it! ... I officially completed my undergraduate program and received my bachelors degree. may the glory be to God for blessing me with the gifts to achieve this great milestone
- Not Frustrated
Of 1.1M tweets tagged with needs, our model predicted over 792K tweets expressing frustration. Below are few example tweets marked as frustrated.
Here’s the week wise transition in the volume of frustrated tweets expressing basic and advanced needs. There is a huge jump in the volume of both categories of tweets. More tweets expressing frustration due to basic needs were posted during the lockdown in comparison to the second wave. The volume of basic tweets during the first wave remained slightly above the pre-COVID level.
The proportion of frustrated tweets across basic and advanced level of needs is illustrated below.
Almost 80% of tweets expressing basic needs are unmet irrespective of the time of the year. It seems that users discuss basic needs only when these needs are unfulfilled. The general rate of frustration for advanced needs is 60%. As soon as the frustration due to basic needs reduces, the frustration due to advanced needs increased by over 10%.
We used VOSviewer to learn the themes of frustrated tweets posted during lockdown and the second wave of COVID-19.
The above figure illustrate the key concerns expressed during Lockown. Travel concerns due to the imposed nationwide lockdown are evident from the terms in cluster blue as shown above. Major Indian cities namely bengaluru, bihar, pune coupled with transportation choices such as bus, train, vehicle can be seen. There are terms such as quarantine, doctor, patient, office in the same cluster indicating the traveling problems faced during daily life activities. The nodes in green reveal the challenges faced by logistics and travel industry. Terms such as refund, ticket, airline, flight, credit reflect the chief complaints by customers along with bill and other payments.
The nodes in cluster red highlight the discussion on digital media and news channels. Growing concern due to increasing toll of infections in the USA and a sense of anger towards China were expressed through tweets. Fake news, channel, minority and economy were also a few topics of online discussion. The nodes in cluster yellow depict the concerns revolving around closed educational institutions, payment of fees, online classes *and *exams.
The below figure illustrate the key concerns expressed during the Second Wave.
The usual customer care complaints are depicted in cluster green. The nodes in cluster red particularly reveal the frustration against political parties and elections. There are also terms such as player, season, ball, game due to upcoming IPL cricket matches. The anxiety due to shortage of ventilator, patient, hospital, icu bed * and *oxygen cylinder is captured through nodes in cluster blue. Words such as refer, friend, help reveal anxious attempts to locate healthcare through contacts on Twitter. Availability of vaccine and booking of slots were also a cause of frustration amongst Indians. Education remained a concern during the second wave as evident from nodes marked in purple.
A minimally supervised approach to annotate tweets with their need level as in Maslow’s Hierarchy of Needs can greatly reduce the time and human effort without much impact on the quality of annotation. A recurring pattern in the needs, indicating predictability in the emerging needs was observed. Pretrained language models used for detecting unmet needs detection, generalises well and is capable of transfer learning from previously labelled data at the start of a crisis.
[1] https://ai.facebook.com/blog/roberta-an-optimized-method-for-pretraining-self-supervised-nlp-systems/
Read more: Rai, S., Joseph R., Thakur P.S. & Khaliq A.M. (2022). Identifying Human Needs through Social Media: A study on Indian cities during COVID-19. Social NLP, NAACL 2022 . Available at Link
]]>
Consider the above mapping where the elusive notion of DREAMS is mapped to a more widely observed entity BUTTERFLIES to express meanings such as vibrant and colorful in a conversation.
Metaphorical constructs are everywhere; in our daily discussions, drama, political speeches, and poetry. These constructs reflect one’s inner feelings and are used to guide readers to intended meaning. Look at these political phrases such as khooni panja for congress’s election symbol, maut ka suadagar, Upadravi Gotra and many more crafted carefully to deliver the target message amongst masses.
When COVID-19 striked India, metaphors comparing COVID-19 virus to enemy, monster and even devil were widely used. Pandemic was equated to war, tsunami, race etc. The aim was to attract people’s attention, convey the urgency of the situation and infuse a feeling of togetherness. However, the discussion quickly took an unexpected turn where these mapping actually started creating a rift amongst different sections of society. Metaphors borrowed from the domain DARKNESS, DISASTER, and STIGMA led to pessimistic perception with a nagging sense of helplessness and isolation. So, how do we track people thoughts, feelings and reactions when the meanings are not explicit and expressed using figurative constructs?
Diachronic word embeddings are used to track semantic shifts in the meaning of a concept. We can thus model the shift in the meaning of metaphorical constructs by analysing embedding space made from tweets in discrete time space namely,
Below is the first plot illustrating the neighbourhood of covid during the lockdown. The representations are learned using word2vec skipgram model by training on metaphorical tweets posted during the lockdown phase. The words closer to covid indicates high relatedness with covid-19.
Here, COVID-19 is conceptualized as WAR (fight, battle, deadly, break, war), MONSTER (protect, threat, deadly, dangerous), and even as an OBSTACLE/GAME (challenge, overcome, strike, tackle). India had only few reported cases of COVID-19 in the lockdown phase. Nevertheless, the metaphors are grim and fear inducing.
Post-Lockdown is the phase where India faced the first wave of COVID-19. COVID-19 is discussed using concepts such as race, win, fight, combat, tackle. Moreover, even WAR metaphors are used in more authoritative fashion indicating the transition in the meaning of WAR metaphors while discussing COVID-19. Metaphors such as duty, strategy, push, tackle, brave, fighter, crusader convey a sense of control, a more confident and controlled reaction to the pandemic.
Highest volume of metaphors were posted in the Second Wave phase. Negative metaphors such as battle, deadly, crisis, suffer, defeat, dangerous came closer to covid compared with the previous two phases. Increased occurrence of DISASTER/DARKNESS metaphors such as grim, catastrophe, devastating, disaster, panic, nightmare, gloom, danger, tsunami etc is observed. Metaphors such as breach, blame, rage, pressure from GAME domain indicating the shift in meaning of GAME metaphors.
There is a clear difference in the underlying emotional tone of metaphors when compared with the first wave phase. The first wave definitely saw a controlled strategy with manageable COVID-19 cases whereas the second wave witnessed more suffering, panic and lack of control which is also evident from the metaphors.
Metaphors serve as the mirror of one’s psyche and are helpful in understanding the pulse of a society. Our experiment revealed that source domains such as WAR, MONSTER, PRISON, LESSON/TEACHER, SUCCESS/ CHALLENGE imparted a more positive influence in the Covid-19 discourse on twitter. This reaction of people is in contrast to prior work who state that the use of WAR and related metaphors paints a more negative picture about the pandemic. Looking through example tweets from these domains in an Indian context, we see that the presence of such metaphors excites a feel of encouragement on Twitter wherein Indians feel it as a duty to fight and defend their country from COVID-19.
Khaliq, A M., Joseph, R. & Rai, S. (2021) #covid is war and #vaccine is weapon? COVID-19 metaphors in India. In the 18th Proceedings of International Conference on Natural Language Processing (ICON), NLPAI. Available at Link
Image credits: Shutterstock
]]>