Winter Storm 2024 will be January 9th to 12th and January 16th to 19th

Winter Storm is the UMD Language Science community’s yearly two-week series of seminars and workshops designed to foster research skills, stimulate new interdisciplinary research projects, and boost career growth and job readiness. These workshops are FREE and open to all language scientists, encompassing undergraduates, research assistants, grad students, postdocs, faculty members, and researchers at affiliated research institutes.

RSVP to Winter Storm 2024 here!

Program Details

Check out the Winter Storm 2024 schedule here!
Click each workshop title to learn more:

Technical Skills Workshops
Introduction to Programming in R, Kathleen Oppenheimer (HESP)
LLaMa Tell You a Secret: How to Work With LLMs as Task Assistants, Nishant Balepur (CS) and David Martinez (ARLIS)
Intro to pcIBEX for Experiments, Lauren Salig (NACS)
Data Visualization Using R and Adobe Illustrator, Dushyanthi Karunathilake (ECE)
Data Preprocessing & Cleaning, Zhiyi Wu (SLA), Nan Zhang (HDQM), Lauren Salig (NACS)
Multilevel Modeling for Psycholinguistic Data, Zhiyi Wu (SLA), Nan Zhang (HDQM)
LaTeX & RMarkdown: Because Who Needs a Social Life Anyway?, Cassandra Caragine, Jessica Mendes, Utku Turk, Malhaar Shah (LING)
LLMs in Language Science Research, Sathvik Nair (LING), Rupak Sarkar (CS), Ishani Mondal (CS), Zongxia Li (CS)

Research Workshops
Hot Topics in Sign Language Research I: Children Learning Language by Ear and by Eye (Bimodal Bilingualism), Deborah Chen Pichler (Gallaudet)
Using Online ASL Resources for Research, Deborah Chen Pichler (Gallaudet)
Hot Topics in Sign Language Research II: How Modality Impacts Second Language Learning, Deborah Chen Pichler (Gallaudet)
Investigating Neural Speech Processing Using MEG and Python, Dushyanthi Karunathilake (ECE)

Professional Development Workshops
Language Careers in Tech Companies: Perspectives From Former Terps, panel moderated by Shevaun Lewis
Career Planning in Graduate School, Yi Hao, Program Director of Career and Professional Development, The Graduate School
Career Planning: Get Started!, Shevaun Lewis (LSC)

"Nobody's Doing This" Lightning Talks
January 9th
"If It’s Not Stuttering, What-Ut-Ut-Ut Is This-Is?" by Vivian Siskin
"Real-Time Measurement in Spoken Discourse" by Kelly Marshall
"Your Native Language Matters in Writing a New Script" by Xianglin Zhang

January 12th
"Secrets of the Internal Writing Sequence for Language Learners" by Yi Dai
"Gaps in Developmental Speech Perception Research" by Rochelle Newman
"Why Isn't Phonics Put To Work for Me? (And Other Adult SLA Oversights)" by John Scott (Scott will also hold a research/discussion group) 

January 17th
"Can Children Learn From Overheard Speech? If So, Who and How?" by Rachel Romeo
"We Don’t Know How the Brain Stores Anything, Let Alone Words" by Bill Idsardi
"Place, Platform, and Presentation of Personality" by Anton Rytting

Workshop Descriptions

Technical Skills Workshops
Introduction to Programming in R, Kathleen Oppenheimer (HESP)
Location: 2123 HJ Patterson Hall
This intro programming workshop is designed for programming novices and R users who are not yet comfortable writing their own code from scratch. Those with experience programming in other languages who are new to R are also welcome. I plan to cover the following topics and will adjust as needed based on the interests of the group: using and writing functions, cleaning and reshaping data, basic data visualization, debugging, and finding resources to continue your learning outside of the workshop. We will start off using a simple dataset that I provide, and then we’ll move on to apply our programming skills to real language science datasets. Attendees are encouraged to bring their own data to work on but it is not necessary.
Before the first session, attendees should install R and RStudio. They should bring their computer to each session. If these do not work on your computer for any reason, or if you are not able to bring a computer, please contact Kathleen via email (koppen@terpmail.umd.edu) or via Slack before the first session so we can figure something out.

LLaMa Tell You a Secret: How to Work With LLMs as Task Assistants, Nishant Balepur (CS) and David Martinez (ARLIS)
Location: 2124 HJ Patterson Hall
Do you have data that needs to be manually and perhaps subjectively coded (e.g., data cleaning and preprocessing, evaluating responses to open-ended questions)? This workshop will provide an introduction to LLMs as task assistants. Attendees will learn the fundamentals and best practices of LLMs such as ChatGPT and LLaMA (Meta’s open-source LLM similar to GPT-3) and how to work with them to effectively and efficiently complete natural language tasks.
The workshop will include brief lectures to introduce key topics; hands-on demos; as well as plenty of time for attendees to work with their own data. To maximize the benefit of this workshop, we strongly encourage attendees to bring their own data, though we will be able to provide at least one data set for those who need it. For this workshop, the ideal dataset should lend itself to a natural language task, in English, of the kind that would normally be given to RAs, annotators, or the like. (For other kinds of tasks, e.g., topic modeling, see the workshop titled “LLMs in Language Science Research”.) We will interact with LLaMA via Python. Attendees should have some experience coding in a computer language (e.g., R) and will need some familiarity with Python, however, the first day of this workshop (“Day 0”) is entirely devoted to introducing Python. Those who have at least a basic familiarity with Python can begin joining us on the second day of this workshop—if you can import a dataset in Python and find the mean of a vector, then you can probably begin joining on the second day. Note: attendees will be free to use the programming language of their choice to prepare and analyze their data (e.g., to convert data to long format).

Intro to pcIBEX for Experiments, Lauren Salig (NACS)
Location: 2123 HJ Patterson Hall
Are you searching for a good way to run psycholinguistic experiments and considering PCIbex? Come join me! In these sessions, we’ll walk through the basics of setting up an experiment on PCIbex, discuss how to create your specific studies, and briefly cover how to extract and analyze data from PCIbex. Please create a free PCIbex account here before our first session. You may also want to walk through the PCIbex Basic Tutorial to have a sense of how it works, although this is not strictly necessary. Click here for a basic tutorial.
Our second session will be more of a workshop where we cover what YOU want to know to set up your experiments. Especially for that session, please bring your questions, error-ridden PCIbex scripts, study designs, etc.

Data Visualization Using R and Adobe Illustrator, Dushyanthi Karunathilake (ECE)
Location: 2124 HJ Patterson Hall
As researchers and students, we often encounter the challenge of transforming statistical results and concepts into visually compelling presentations, like figures for articles, posters, and presentations. This hands-on workshop aims to guide you through this process. First, you will learn to generate statistical graphics in R using ggplot2 and then, you’ll learn how figures created in R can be refined and polished further using Adobe Illustrator.

Data Preprocessing & Cleaning, Zhiyi Wu (SLA), Nan Zhang (HDQM), Lauren Salig (NACS)
Location: 2124 HJ Patterson Hall
Tired of wrestling with messy raw data? Join our one-session workshop, where we will cover essential R functions to get your raw datasets ready for more advanced data analysis using R studio. There will be step-by-step hands-on practices. Bring your own datasets to get help untangling them! Attendees should bring their computer to the workshop with R and RStudio installed and are strongly encouraged to bring their own datasets.

Multilevel Modeling for Psycholinguistic Data, Zhiyi Wu (SLA), Nan Zhang (HDQM)
Location: 2124 HJ Patterson Hall
Ever wondered how to run a mixed-effects model to address both within-group and between-group variations in your dataset? Join our hands-on workshop, where we will guide you through the step-by-step process of running multilevel analysis (e.g., item level and subject level), examining the assumptions for the models in R, all tailored for psycholinguistic data analysis.

LaTeX & RMarkdown: Because Who Needs a Social Life Anyway?, Cassandra Caragine, Jessica Mendes, Utku Turk, Malhaar Shah (LING)
Location: TBA
Are you a fellow procrastinator who finds solace in tweaking your LaTeX code? Isn't coding in R annoying on its own? Well, join us and discover how we can assist you with both. We'll be covering our favorite packages for glossing linguistic experiments, along with handling phonological, syntactic, and semantic typesetting. Later, we'll demonstrate how to integrate them into an RMarkdown/RSweave (or Quarto) environment, which can help you incorporate your R or Python code to generate submission-ready PDFs.

LLMs in Language Science Research, Sathvik Nair (LING), Rupak Sarkar (CS), Ishani Mondal (CS), Zongxia Li (CS)
Location: 2123 HJ Patterson Hall
Session one: we will use language models as a tool to help us analyze data from psycholinguistic experiments. Previous generations of LLMs are far more transparent than ChatGPT, and can explicitly provide us with quantitative measures that we can use to work with behavioral or neural data. We will present approaches on how to compute the similarity between words and score the complexity of sentences in stimuli for experiments, which can then be compared with human responses. If possible, click here to download the Python library minicons. Those who do not download Python library minicons can do so during the session. In this session, we will also reserve some time for applying these techniques to your data.
Session two: Large Language Models (LLMs) such as ChatGPT, Claude, and others are gaining more popularity as collaborative tools for various professional and personal purposes. These AI models offer a wide range of assistance such as writing, coding, content analysis, document interpretation, image interpretation, question answering, etc. While these models are designed to adapt to a diverse range of users for any purpose, they encounter limitations in specialized or personalized tasks. Example tasks like assigning labels to professional documents and analyzing specialized passages with expert-level precision often reveal the general purpose LLMs shortcomings. LLMs still lack the expertise to produce ideal outputs with specific requirements. We introduce the concept of in-context learning and fine-tuning, which are two useful and easy ways made accessible by OpenAI, to make LLMs suit your specific task requirements. Managing large volumes of text data solely with LLMs presents challenges, as these models have constraints on input text length and demand significant computational resources for extensive data analysis. Such resources may not be readily available to all users. We introduce the Topic Model, designed to automatically uncover semantic themes within a large corpus of unstructured text. This method efficiently organizes texts into distinct topics, grouping passages that are semantically similar or related, thereby facilitating easier data analysis.

Getting Started (Or Getting Caught Up) in Zotero, Alex Krauska and Katherine Howitt (LING)
Location: 2123 HJ Patterson Hall
Reference management software is a great way to stay organized while doing research, but it can take a lot of work to get started and to keep up with it. Join us for a quick crash course on getting started in Zotero, and then stick around to work on organizing (or re-organizing) your library. This session is open to beginners and experienced users of Zotero alike!

Research Workshops
Hot Topics in Sign Language Research I: Children Learning Language by Ear and by Eye (Bimodal Bilingualism), Deborah Chen Pichler (Gallaudet)
Location: 2114 & 2118 HJ Patterson Hall
What happens when babies learn both a sign language and a spoken language as their first languages? How do the two languages interact, and how do those interactions compare with the more familiar context of bilingualism in two spoken languages? These questions motivate current research on bimodal bilingualism, or bilingualism across two different modalities (signed and spoken), and lead to unique insights on how human minds acquire, organize, and process more than one language. They also lie at the heart of heated debates about language choice and early intervention practices for deaf and hard of hearing (DHH) children. In this talk I will draw from the growing body of research on bimodal bilingual children, including studies that I have been involved in over the last 20 years, to summarize notable insights about language development revealed by this research.

Using Online ASL Resources for Research, Deborah Chen Pichler (Gallaudet)
Location: 2123 HJ Patterson Hall
Sign language research has broadened our understanding of how humans develop as multilinguals in both spoken and signed modalities. Today researchers have access to an unprecedented volume of sign language data in the form of online videos, corpora, and other databases. This talk highlights research that has been carried out using popular online sign language resources such as ASL-LEX (https://asl-lex.org/) and ASL Signbank (https://aslsignbank.haskins.yale.edu/). 

Readings:
Sehyr, Z. S., Caselli, N., Cohen-Goldberg, A. M., & Emmorey, K. (2021). The ASL-LEX 2.0 Project: A database of lexical and phonological properties for 2,723 signs in American Sign Language. The Journal of Deaf Studies and Deaf Education, 26(2), 263-277. https://doi.org/10.1093/deafed/enaa038
Caselli, N. K., & Pyers, J. E. (2017). The road to language learning is not entirely iconic: Iconicity, neighborhood density, and frequency facilitate acquisition of sign language. Psychological science, 28(7), 979-987. https://doi.org/10.1177/0956797617700498

Hot Topics in Sign Language Research II: How Modality Impacts Second Language Learning, Deborah Chen Pichler (Gallaudet)
Location: 2114 & 2118 HJ Patterson Hall
In part two of "Hot topics in sign language research," I will turn to (adult) second language learners of sign languages. Again, the overarching question is how modality affects language development, but this time in the context of sequential bilingualism. To date, the bulk of research in this area focuses on hearing high school and college students learning their first sign language. These learners exhibit many of the familiar L2 patterns from spoken language acquisition, as well as some unexpected patterns that are assumed to be unique to L2 acquisition in a new modality. However, such assumptions may be premature in the absence of a crucial comparison group: deaf signers learning an additional sign language. This talk overviews major findings about L2 hearing learners of various sign languages and points out the importance of extending this research to deaf multilingual signers to achieve a more comprehensive understanding of modality effects on language acquisition and bilingualism.

Investigating Neural Speech Processing Using MEG and Python, Dushyanthi Karunathilake (ECE)
Location: 2124 HJ Patterson Hall
Are you interested in understanding how the brain comprehends speech using magnetoencephalography (MEG) data recordings? In this research group, we will discuss recent MEG studies that explore the temporal dynamics of continuous speech processing across the speech and language hierarchy. In addition, you will gain proficiency in analyzing MEG data- from pre-processing to source localization, and temporal processing of different speech features using Eelbrain and MNE Python package

Toward a 21st-century Adult SLA Pedagogy, John Scott (SLLC)
Location: 2123 HJ Patterson Hall
What could/should L2+ language instruction for adults look like in the rest of the 21st century? After decades of working largely in parallel but not as often across disciplines, various fields including Linguistics & SLA, Hearing & Speech, Education & Early Literacy, and Applied Linguistics have investigated different aspects of language learning in an increasingly plurilingual and literate world. Each field has made major advances in our understanding of the processes, needs, challenges, and environments of instructed adult L2+ acquisition. Meanwhile, as the social and institutional contexts for foreign language instruction rapidly change, arguably too few of these advances reach the beginner classroom or even professional training for language educators. What do we know? What questions are we still working on, and what might we be capable of doing in adult foreign language instruction to demonstrate our continued relevance in a world that increasingly places universities and colleges in competition with much cheaper private sector alternatives? In this roundtable discussion, we’ll share some of the insights each of our fields brings to the table, brainstorm questions to explore further together, and ways we could bring what we do know to bear for the sake of L2+ learners.
 

Professional Development Workshops
Language Careers in Tech Companies: Perspectives From Former Terps
Location: 2114/2118 HJ Patterson Hall
Panel moderated by Shevaun Lewis
Tech companies offer a range of roles for people with expertise in language, even those without a technical background. In this panel, you'll hear from three former Terps who are all currently working on different teams at Amazon: Jon Burnsky (BA ‘16), Aaron Doliana (PhD ‘21), and Tess Wood (former LSC Assistant Director).

Career Planning in Graduate School, Yi Hao, Program Director of Career and Professional Development, The Graduate School
Location: 2114/2118 HJ Patterson Hall
Career planning shouldn’t be an intimidating task that you put off until your last year in graduate school. In this session, you’ll learn about how to fit career development into your academic program timeline, and how to use resources available from UMD to research potential careers, set short- and long-term goals, and eventually search for jobs.

Career Planning: Get Started! Shevaun Lewis (LSC)
Location: 2124 HJ Patterson Hall
If you’re feeling overwhelmed at the prospect of figuring out what career is right for you, try these exercises for self-assessing your interests, skills, and values.

"Nobody's Doing This" Lightning Taks
Location: 2114/2118 HJ Patterson Hall
"Nobody's Doing This" Lightning Talks are 3-5 minute talks that cover and discuss the development of new research ideas. Speakers address glaring gaps in research, projects in the back of their minds, or ideas that they haven't had time to revisit... until now! 

Lunch & Coffee
Location: 2130 HJ Patterson Hall

Social
Location and details TBA

Schedule adjustment: Due to the delayed opening on January 10th, Intro to R and Working with LLMs will be from 10:00 to 11:25 AM, Hot Topics in Sign Language: I will be from 11:30 to 12:30, and lunch will be served at 12:30,

Schedule