Join the Text Analytics Working Group!
The text analytics field (also known as Natural Language Processing or NLP) has been one of the fastest growth areas of AI capability since 2018. NLP is the ability of computers to understand and draw insights from text and human speech. Recent advances have been driven by new deep learning (transformer) models which radically improve the power of text analytics tools.
New text analytics approaches offer huge opportunities for MERL practitioners to transform M&E systems by enabling processing and analysis of vastly more data (for example, that which lives in PDF documents) in less time at reduced cost. MERL practitioners can also enrich their methodological toolkits through computer-assisted automation.
The MERL field has been slow to understand and adopt NLP technologies as compared to other fields, such as business intelligence and market research. Yet MERL practitioners are urgently needed in debates over the ethical, inclusive and responsible use of NLP technologies.
To increase the knowledge, skills, and influence of MERL practitioners in the area of NLP, MERL Tech is kicking off a Text Analytics Working Group. Our first (virtual) meeting will be on January 19, 2023 at 8am ET time/1pm UK/4pm Nairobi time. Register at this link to join us. Upon registration, you will receive a link for the call: https://zoom.us/meeting/register/tJUpcOyprjwuG9xkNl0IPy5zTVt8g8ShA9_d
Objectives of the working group
- Raise awareness and understanding about what text analytics (NLP) technologies can do for MERL and how to use these technologies ethically and responsibly, especially in Global South contexts
- Strengthen networks between MERL practitioners (including data scientists) to assist with the uptake and scaling of NLP technologies through the co-development of tools, systems and sharing of open data and code
Who should join?
The working group welcomes all Monitoring, Evaluation, Research and Learning (MERL) practitioners and data practitioners interested in using text analytics approaches for MERL in the development, humanitarian, peacebuilding, and related social sectors. All MERL practitioners are welcome, no prior knowledge of text analytics is needed. Specialist data scientists and NLP specialists are also welcome. Working Group meetings will cover both technical and non-technical aspects; e.g. ethical and inclusive use. We especially encourage practitioners living and working in the Global South to join.
Working Group session format
The 1-hour first session will be facilitated by Matthew McConnachie (Niras), Jonas Noren (Dav|Con) and Linda Raftree (MERL Tech). We will start the session with a presentation and group discussion about the overall applications of text analytics for MERL and ethical considerations. Jonas will then lead a presentation on approaches for extracting web-data, followed by a discussion on this topic.
The frequency of the working group meetings will be decided during the first session; it will likely be quarterly or every second month.
During the first session we also will discuss with group members preferences for structuring future meetings. One idea could be for each session to focus on a specific text analytics topic, facilitated by a volunteer working group member.
We will be creating a Slack channel to support ongoing collaboration and information sharing between the group members. Other channels or work spaces might be created as well, such as LinkedIn, GitHub or Hugging Face if the group should produce code or open models for sharing.
Register now! https://zoom.us/meeting/register/tJUpcOyprjwuG9xkNl0IPy5zTVt8g8ShA9_d.
Leave a Reply
You might also like
-
What’s happening with GenAI Ethics and Governance?
-
Join the AI and African Evaluation Working Group Meet ‘n’ Mix Session on May 7!
-
Hands on with GenAI: predictions and observations from The MERL Tech Initiative and Oxford Policy Management’s ICT4D Training Day
-
When Might We Use AI for Evaluation Purposes? A discussion with New Directions for Evaluation (NDE) authors
Hi Linda
Because of time zone differences I cant join this meeting. But I would like to join the group and be included in any mailing list.
ParEvo.org generates lots of text per exercise, around 10 to 12 thousand words per exercise. I have used topic modeling on one set of text data, and would be interested to hear others views on its utility
Hi and thanks, Rick! I’ll make sure you get on the mailing list. And I’m happy to ask the group about ParEvo.org as well. -Linda
Because of time zone differences I cant attend this meetings. But I would encourage members to pay special attention to the text analytic capacities of ChatGPT, developed by OpenAI and now open to general users. I tried it out as a means of doing three types of analysis of storylines developed during a parevo.org exercise. Firstly, I asked it to summarise in 300 words, the text of one complete storyline, about 1200 words long. Then, I asked it to identify the most noticeable differences between three storyline endings. The I asked it to identify people, groups, organisations and nations mentioned in three storyline endings. The results:
Task 1. This was done well but I wondered if it would produce a very similar summary for the other storylines, i.e it was not picking up on the unique features of the storyline. Hence the second task
Task 2: This was reasonably done well, but I was looking for a more set theory description of findings i.e this one has x, but the others do not, this one has y but the others do not
Task 3: This was done very well and so much easier than finding and using specialist software for noun extraction
The important point to note about ChatGPT is that it works on a dialogic basis. You ask it to do a task, it respond, you ask it to refine that product, it responds, you ask it to a further refinement, etc. I did not explore this feature, which could have improved results on task 1 and 2. You can also ask it to have multiple repeat attempts at the same task
From this exercise and what I have read and seen on videos about ChatGPT, the capacities are genuinely awesome, though with some recognized limitations, and of course potential for misuse. This is one video I watched as an intro
https://www.youtube.com/watch?v=40Kp_fa8vIw&t=147s and another by same person
https://www.youtube.com/watch?v=JTxsNm9IdYU&t=29s
Thanks Rick – I’m definitely interested in digging more into the pros and cons and utilities of ChatGPT as one of the tools that we take a look at!
Hi Linda,
I’m sorry I missed this first session. I learned about this new MERLTech group through AEA. Could you please add me to the mailing list?
Will add you – check your inbox!
Hi Linda,
I’m afraid I missed this session but I’d love to join the group, go on the mailing list and come to the next one! Would that be possible?
Absolutely! I’ll add you to the mailing list.
Thank you!
i also missed the meeting, can i register now please and someone can add me in the mailing list
Added you to the list!
Hi, I missed this session and interested in joing the working group. Could I still be added to the group?
Have added you – check your inbox!
Hi! I missed the meeting as well, but I’d love to be added to the mailing list for the working group. Many thanks!
Hi Linda,
I missed the first session. Kindly requesting you to add me to the mailing list. Thanks
Dear Linda,
I missed the first session, but would love to participate in the future. Could you please add me to the mailing list?
Best wishes,
Andrea