Discover and read the best of Twitter Threads about #NLP

Most recents (24)

🚨 New paper! 🚨 w/ @william__brady & @manos_tsakiris:

Today in @PNASNexus we map the moral language of 39 U.S. presidential candidates to show how they're connected / differentiated by their use of moral rhetoric.
ow.ly/8Zwu50OO5QA

A [visual] thread on findings: Image
Main takeaways:

1) @TheDemocrats & @GOP
candidates use sharply divergent moral vocabularies

2) Candidates can separate themselves from the rhetorical norms of their party by using unique moral language, but rarely do
Taken together with prior research (@RobbWiller), this work empirically shows that the moral rhetoric employed by recent candidates is probably un-persuasive across party lines & likely exacerbates polarization, particularly in online social networks.
Read 13 tweets
1/ 🧪🔬 Ever wondered if scientific research can be predicted? Our latest paper "What's next? Forecasting scientific research trends" tackles this question. Check out the details here 👉 arXiv:2305.04133v1 [thread] 🧵
2/ 🌍 The world of science is constantly evolving. But what if we could predict these trends, especially in life sciences? That's what our work aims to do. Imagine the possibilities!
3/ 📚 We've mined historical publications, research/review articles, and patents. The result? Models that can foresee scientific trends 5 years in advance. Time travel, anyone? #AI #DataScience #timeseries #trends
Read 11 tweets
Welcome to #chatbotvoice2023 from Berlin.

I'm throwing up some tidbits around #conversationalai live from the 5th World Chatbots & Voice Summit from Berlin.
First up, Christoph from @GetCyara

We're talking testing and assuring quality in #conversationalai Image
Testing is an impossible manual task and it needs to be automated. Done wrong, it can cost heavily. So it is a huge, but important challenge. Image
Read 83 tweets
🧑🏽‍⚖️ Prediction: supreme courts in many countries will someday debate scraping #data to train models without permission.

🧵 1/5
Many of the #NLP (natural language processing) and #LLM (large language models) have not asked permission of the content creators to use their content.

2/5
Some of the litigation is already happening. See the interview @nabihasyed did with Katherine Forrest for @themarkup: themarkup.org/hello-world/20…

3/5
Read 5 tweets
🧵Hugging Face, a leader in the #AI community, has revolutionized the creation, collaboration & discovery of ML models & apps.

By open-sourcing their tech, they've become a force in the industry, raising $160M & serving 10k+ companies.

Here's why their approach is key! #ML #NLP Image
🚀Open-sourcing allowed Hugging Face to foster collaboration among researchers & companies, propelling the launch of #ChatGPT - a tipping point in AI's shift from niche to mainstream. #OpenSource #MachineLearning
🌐With a focus on natural language processing & text-to-image tasks, Hugging Face expanded to other domains, hitting 50M page views in Jan.

Open science & open source are vital for ML advancement, despite some companies' commercial goals. #NLP #ComputerVision
Read 13 tweets
🚀 1/ Excited to share our (with Aydar Bulatov and @yurakuratov ) report on scaling Recurrent Memory Transformer to 2M (yes, two millions)😮 tokens! 🧠🌐 #AI #NLP #DeepLearning Image
2/ 📈 We've tackled the quadratic complexity of attention in #Transformers by combining token-based memory & segment-level recurrence, using RMT.
🔸 RMT adapts to any Transformer family model
🔸 Memory tokens provide the recurrent connection 🎛️💡 #AI #NLP #DeepLearning Image
3/ 🧠 We tested RMT's memorization capabilities with synthetic datasets requiring fact memorization, detection, & reasoning. The model must separate facts from irrelevant text and use them to answer questions in a 6-class classification. 🎯 #AI #NLP #DeepLearning Image
Read 10 tweets
Revolutionizing the World: 20 AI & Machine Learning Startups You Need to Know. 🧵...
1/20: In this thread i will try and explore the fascinating world of AI and machine learning startups! Discover some of the most innovative companies pushing the boundaries in this space. #AIStartups #MachineLearning
2/20: First up is @OpenAI, the team behind the groundbreaking GPT series.

With GPT-4, they're developing even more advanced natural language processing capabilities to revolutionize human-computer interactions. #NLP #OpenAI
Read 22 tweets
1/Behavioral Health Link develops & markets #CrisisLines services to #988Lifeline providers. One of their offerings: "Voice-Currently in R&D, voice analysis technology to provide real time feedback on a caller’s emotional state to improve care delivery."web.archive.org/web/2022120220…
2/I've started mentioning Dr. John Draper's connection there, recently hired as their President of R & D. You may recall Draper was the long-time leader of Vibrant Emotional Health's administration of the Nat'l Suicide Prevention Lifeline,now #988Lifeline. linkedin.com/in/john-draper…
3/Between end of March & now, Behavioral Health Link took down its page of software solutions. Maybe it will go back up but I'd like to know why it's gone. Seems they've sanitized the site of their " #AI " type services. No doubt still in development. web.archive.org/web/2023041614…
Read 11 tweets
Le gouvernement (service PEReN) a publié un document pédagogique décrivant comme fonctionnent les IA conversationnelles comme ChatGPT #GPT4, en voici une synthèse rapide. 1/9 Image
🧠 Les Large Language Models (LLM) sont d'impressionnants réseaux de neurones qui révolutionnent le Traitement Automatique des Langues (TAL ou NLP). Ce sont des modèles statistiques qui imitent le fonctionnement du cerveau humain pour traiter et comprendre le langage. #NLP 2/9
🏗️ Les réseaux de neurones sont structurés en couches, chaque couche détectant des motifs récurrents (patterns) dans les données. Les neurones contiennent des paramètres qui sont ajustés au cours de l'entraînement pour améliorer les performances. #MachineLearning 3/9 Image
Read 9 tweets
Wie mittlerweile wohl alle mitbekommen haben, wurden die Themen #Bevölkerungsschutz, #Katastrophenschutz und #Zivilschutz eher stiefmütterlich behandelt. Wir wurden von einer Pandemie auf dem ganz falschen Fuß erwischt, im Ahrtal haben viele Menschen Existenz bzw. ...
...sogar ihr Leben verloren. Und plötzlich merkt man dann, dass ein Irrer aus dem Osten jetzt völlig hohldreht und die Landesverteidigung schon fast als marode bezeichnet werden kann.

Alles wurde zurückgebaut. Die #Heeresflugabwehr wurde abgeschafft. Deren "Spielzeug" Gepard ...
... kann sogar Dauerfeuer schießen 😎 und ist dafür sogar gemacht. Es gab deutschlandweit Sprengschächte in Brücken, um diese im Fall der Fälle kontrolliert unbrauchbar zu machen ohne sie komplett zu zerstören. Und, und, und...
...
Read 18 tweets
Interesting failure mode of #GPT4!

It can't play "Set", a card game that is trivial to solve with a 10-line python program.

--> it can explain the game, can abstract it, write a program to solve it, but can't actually *play* it.

Check the full convo in the chat below:
First, I asked about the game (great intro if you don't know the game).
Then, I asked #GPT4 to come up with a strategy to solve it. The strategy is sound, but it doesn't consistently map between the numbers and the attributes:
Read 9 tweets
I recently experimented with OpenAI's API using LangChain. The possibilities of what we can achieve with this technology are truly endless! In this thread, I'll share how I used LangChain to build an app that generates responses to questions based on YouTube video content. Image
First, I extracted the transcript from the videos. For longer videos, I split the document with enough overlap(to ensure continuity), sent the smaller parts to the LLM, got the summary for each part, and combined them for the final result. #OpenAI #LangChain #AI #NLP
Next, I used LangChain's map-rerank method to get the final answer. By tuning the prompts, I was able to generate responses limited to the video's context. #AI #NLP #MachineLearning #ChatGPT
Read 4 tweets
Natural Language Processing is a Military Intelligence tool used in automated decision making. #NLP
Natural Language Processing:

Security- and Defense-Related Lessons Learned

rand.org/content/dam/ra…
I have been using NLP in SEO/ Content Marketing for years. NLP is not new. It is not good or bad. I am just trying to help people understand the history of NLP because I think it is important to know.
Read 4 tweets
"Ain, Fê: é muito difícil alguém ser demitido no Japão! A #Google cuida das pessoas... ". Cuidava, Fia. Cuidava... O Google Japan acaba de mudar essa realidade. Vem passaralho japonês aí.

#layoff #demissão #demissãoemmassa #recolocação #techlayoffs

asahi.com/articles/ASR32…
Em 02/03/2023, o gigante tech, em profunda reestruturação global e tocando #layoffs pelo mundo afora, anunciou um pacote de #demissão voluntária, oferecendo "benefícios" para quem escolher sair da empresa até 16/03.

Obviamente, a proposta causou revolta nos squads.
Os funcionários decidiram criar grupos de trabalho para contestar a decisão e forçar a empresa a cumprir a legislação trabalhista. Vão tirar até o último centavo a que tem direito.
Read 4 tweets
Developing a Federal #AI Standards Engagement Plan🧵
The Former Director of Twitters Responsible Machine Learning META division worked at Accenture.

Let's review some highlights from the below doc:

nist.gov/system/files/d…
"We encourage the United States government to adopt a similar whole-of-government approach" Image
Read 15 tweets
You've probably asked yourself what these 5 terms mean...

Analytics, AI, SaaS, Big Data, NLP.

If you work in #SEO, this thread is for you.

Let's explain them like we are 5 years old 🧵
Analytics is the discovery and communication of patterns and insights in data.

Analyzing Google Search Console data is such an example.

This is where many SEOs fall short, as this discipline isn't just filtering data!
Data Science is the parent of Analytics.

You may also want to check this thread to broaden your knowledge.

Read 13 tweets
Data Skills and Analytics for #SEO essentials.

What you need to combine 2 of the most lucrative skills out there.

💯 Updated version with new content.

My 30 best threads in a single place:
Introduction to #nlp to understand how modern search engines work.
The Best SEO Python Libraries You Have to Use.

What I list here is more than necessary to start getting results.
Read 36 tweets
Think Bigger. Where your focus goes, your energy flows. The greater the focus the more clarity you get on your vision and mission.

#clarity #vision #mission

🧵1/5
“I have no motivation in these times” is something I hear way too often. It’s putting the cart before the horse. The truth is that pure, positive, pleasurable motivation comes from having a vision for the future, but it’s up to you to paint this picture.

#vision

🧵2/5
Most people are motivated just enough to avoid the pain of the past. Motivated just enough to not totally fail. If you do this enough, you will fail. Guaranteed.
If you are miserable, try setting a little goal. Maybe having a better daily routine.

#Motivation
🧵3/5
Read 5 tweets
How to get started with machine learning: From 0 to Engineer.

(thread)
1. Math - via @khanacademy

Use these resources (all links in the last tweet):

- Khan Academy - Multivariable Calculus
- Khan Academy - Differential Equations
- Khan Academy - Linear Algebra
- Khan Academy - Statistics Probability
2. Learn Python - via @freeCodeCamp

- Learn Python: Full Course For Beginners
- Intermediate Python: Programming Course
Read 11 tweets
Wanna build the next ChatGPT?

Here's a 6 month learning plan for AI large language models (LLMs) from scratch.

(thread)
1. Familiarize yourself with basic machine learning concepts and Python programming.

Books:

1.1. Introduction to Machine Learning with Python by Andreas Müller and Sarah Guido

1.2. Python Machine Learning by @rasbt
2. Learn about natural language processing (NLP) and its applications.

Books:

2.1. Speech and Language Processing by Daniel Jurafsky and James H. Martin

2.2. Natural Language Processing with Python by Steven Bird, Ewan Klein, and Edward Loper
Read 8 tweets
Creating the perfect ChatGPT prompt?

Follow our 7-step guide for success

A thread 🧵

#chatgpt #NLP
ChatGPT can produce amazing results. But proper ChatGPT prompt construction is essential for accurate and effective output. Follow these 7 steps:

1. Persona simulation
2. Task description
3. Task execution
4. Context/prompt injection
5. Constraints
6. Goal
7. Output format.
1. Persona Simulation

A fictional persona representing a specific group or expertise. In most cases, the chosen persona is an expert in a specific field, but personas can also be customer personas, stakeholders like employees or shareholders, and more.
Read 19 tweets
What is #ChatGPT?

It is all the buzz.. Showing my ignorance here, but I had no idea what "GPT" stood for, I just know I LOVE using it......So I asked ChatGPT.....of course.

Thread:
1/ In it's simplest form the GPT in #ChatGPT stands for "Generative Pre-trained Transformer". It generates human like text. #NLP #MachineLearning
2/ #GPT uses a transformer architecture, which is a type of neural network architecture designed for processing sequential data, such as text.
Read 8 tweets
GM! Here are the top trending repos in web3, curated by top devs in the last 24 hours ✨.

Check the 6 projects in the thread below 🧵👇🏻
Repo openai-cookbook by openai (4565 ⭐️)
+11 new ⭐️ in the last 24h from top devs

Examples and guides for using the OpenAI API

github.com/openai/openai-…
Repo starknet_in_rust by lambdaclass (32 ⭐️)
+6 new ⭐️ in the last 24h from top devs

A Rust implementation of Starknet

github.com/lambdaclass/st…
Read 8 tweets
40 Deep #SEO Insights for 2023:

-In 2022, I told to focus on Natural Language Generation, and it happened.
-In 2023, F-O-C-U-S on "Information Density, Richness, and Unique Added Value" with Microsemantics.

I call the collection of these, "Information Responsiveness".

1/40 🧵.
1. PageRank Increases its Prominence for Weighting Sources

Reason: #AI and automation will bloat the web, and the real authority signals will come from PageRank, and Exogenous Factors.

The expert-like AI content and real expertise are differentiated with historical consistency.
2. Indexing and relevance thresholds will increase.

Reason: A bloated web creates the need for unique value to be added to the web with real-world expertise and organizational signals. The knowledge domain terms, or #PageRank, will be important in the future of a web source.
Read 45 tweets

Related hashtags

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3.00/month or $30.00/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!