The Ultimate Tool for Managing In-House Text Annotation.

Joined April 2018
51 Photos and videos
LightTag retweeted
10 Apr 2022
Great to see new NLP data sets being built with LightTag @LabeledData PrevComp is a labeled corpus for the identification of disease complications, underlying risk factors & prevention measures, for the interaction between between hypertension & diabetes journals.plos.org/plosone/ar…
1
2
4
17 Feb 2022
👇
17 Feb 2022
Labeled data compute = moat
2
16 Feb 2022
We've joined @primer_ai
16 Feb 2022
LightTag is joining Primer! Today, we're thrilled to announce our acquisition of a leading, data labeling platform. LightTag makes it easy to set up labeling tasks for teams of labelers, speeding the time it takes to start training #NLP models. Learn more: hubs.ly/Q014j5rM0
1
4
13 Oct 2021
We’ve been saying this for years
I’m thrilled to share our new work “A Few More Examples May Be Worth *Billions* of Parameters”. Joint w/ @PSH_Lewis @riedelcastro @omerlevy_ Paper: arxiv.org/abs/2110.04374 1/4
4
29 Jun 2021
How can we squeeze out every of goodness from labeled data ? Train on it's complement as well with COT. First proposed by @3F55bCEltngxKyq and team, it's a simple way to boost performance. Here's a primer and how to implement with @PyTorchLightnin lighttag.io/blog/complement-…
1
4
If you love hyper graphs, language and have an incredibly high technical bar reach out to our friends at @codexeditor
30 Apr 2021
I have a Codex announcement: I'm looking for a technical co-founder to help accelerate development & to take Codex to the next level. Familiarity with graphs, full stack, modern JS & UI/UX highly regarded. Experience raising funding, also. Feel free to DM for more details.
1
5
LightTag retweeted
30 Apr 2021
I have a Codex announcement: I'm looking for a technical co-founder to help accelerate development & to take Codex to the next level. Familiarity with graphs, full stack, modern JS & UI/UX highly regarded. Experience raising funding, also. Feel free to DM for more details.
8
31
98
27 Apr 2021
Help us help you label more data faster. Check out this feature we're working on and share your ( constructive) feedback
27 Apr 2021
Trying something new today. I've been working on a feature for @LabeledData and seeking feedback. I have a few #NLProc followers, anyone want to take a look and share feedback ? loom.com/share/14b73e89f6244…
1
3
11 Feb 2021
Tell me you ship NLP to production without telling me you ship NLP to production
1
4
We can make the annotation experience fast and efficient. We unfortunately don't do office furniture
1
3
18 Jan 2021
What has 4 letters, sometimes has 9 letters, but never has 5 letters. That's not a question, read it again. It does show the trickiness of #NLP, sometimes the meaning of a word can change because of a missing question mark far away.
1
15 Jan 2021
#NLP is everywhere, even in finance. This article describes five amazing applications of natural language processing in the financial domain. buff.ly/3s6AmvS

15 Jan 2021
What's an easy way to make your #text processing and #NLP pipelines 10x faster ? Use algorithms from the 60's. In this classic article Russ Cox describes why regular expressions are slow and how with some common sense they can be fast buff.ly/39dCr0o

1
14 Jan 2021
Can #NLP bring down the internet ? Turns out it can if done wrong. Regular expressions, a cornerstone of #text processing, can have subtle but profound performance bugs. One hit Cloudflare and brought down half the internet . Read more in their postmortem buff.ly/2SgCLT5

14 Jan 2021
Everyone in #NLP talks about attention because of the amazing performance of #transformers. But are there other models that could be competitive ? A while back the team at Facebook AI published a paper on "dynamic convolutions" which run faster. buff.ly/3s7YpKH
3
13 Jan 2021
#NLP is good at is #Syntax, figuring out the structure of a sentence. That's pretty useful, we can use syntax to search for structure in text. The team at AllenAI have built a system that figures out the structure you were searching for, with #AI. buff.ly/2JVdsGp

1
13 Jan 2021
How can you choose the right named entity recognition model for your use case ? With so many options in the #OpenSource #NLP world, it can be overwhelming to figure out which one is right for you. Read our article showing how you LightTag can help buff.ly/3rYqMLk
2
1
12 Jan 2021
Human in the loop is one of the most powerful ideas in #AI and #MachineLearning, making our models and our humans more effective. Find out when and how you can use human in the loop to drive your #NLP efforts. buff.ly/3pZb8gX

1
1
12 Jan 2021
#ActiveLearning uses AI to train #AI. Instead of labeling all of our data, we ask our AI what is the highest value data to label, cutting costs and saving time. But Active Learning has some pitfalls that you should know before using it. buff.ly/3bejGfU
1
2