Emotional and Artificial intelligence

Welcome to the 15th edition of the AI and Global Grand Challenges newsletter, where we explore how AI is tackling the largest problems facing the world.

The aim: To inspire AI action by builders, regulators, leaders, researchers and those interested in the field.

Make sure to subscribe to our site as well: www.nural.cc

----

So let's get straight into it!

Packed inside

• Social researchers have built a game allowing you to use your webcam to demonstrate the lack of emotional intelligence of modern AI systems
• The latest on AI and regulation by Nathan Thomas
• A new dataset and corresponding challenge for forecasting localized climate impacts
• and more...

If you would like to support our continued work from ￡1 then click here!

Marcel

__________________________________

Nural Research Article -  AI & Regulation

As the pace of global innovation has undoubtedly been intensified by developments in AI, such seismic changes have inevitably garnered scrutiny from governments and regulators. There is growing fear of a so-called ‘information gap’...

Read the rest

__________________________________

Key recent developments

---

Do as AI say: susceptibility in deployment of clinical decision-aids

What: In a study, physicians received chest X-rays and diagnostic advice, some of which was inaccurate, and were asked to evaluate advice quality and make diagnoses. All advice was generated by human experts, but some was labelled as coming from an AI system.

The results: As a group, radiologists rated advice as lower quality when it appeared to come from an AI system; physicians with less task-expertise did not.

Key Takeaway: There is often little evaluation of the AI- doctor interaction and this study raises some interesting questions. What is the correct balance of trust that doctors should place in the advice produced by AI? In this case, the specialists show a clear lack of trust, but too much scepticism reduces the assistance of the technology.  

---

Demonstrating the flaws of emotion recognition tech through a game

Play the game here: https://emojify.info/menu

What: A group of social scientists have developed an emotion recognition game that allows users to interact with a similar emotion recognition system to that used in AI job interviews, by marketers and by some police. The game demonstrates ways you can trick the system. I recommend trying it out using your laptop webcam!

Key Takeaway: There has been a huge increase in the implementation of emotion recognition systems with job interviews, airport security and court trials just three examples. However, with proven reductions in performance for some demographics and the ability to often mislead these systems as evidenced by the above game, further questions should be explored to ensure they perform as they should.

---

How to Set Up Your ML Pipeline For Post-Deployment Success

What: For those of you looking to implement machine learning into your business or anyone interested in ML deployment. The above checklist provides a very useful framework to manage performance, ethics and post deployment improvements.

Key Takeaway: There are two sides to AI for social good. The social good side is evident, however it is redundant without good ML frameworks for successful deployment!

__________________________________

AI Ethics

?? Alphabet shareholder pushes Google for better whistleblower protections

??How we’re (Facebook) using Fairness Flow to help build AI that works better for everyone

??Artificial intelligence must know when to ask for human help

Other interesting reads

??ProteinGAN: A generative adversarial network that generates functional protein sequences

??An artificial intelligence tool that can help detect melanoma

??Dual Contrastive Loss and Attention for GANs - Improving the performance of image generation by using a new loss function

??Automatically generating text from structured data like tables

Datasets

??SPGISpeech: 5,000 hours of transcribed financial audio for fully formatted end-to-end speech recognition

??Shedding light on fairness in AI with a new data set (Facebook Video Dataset)

??arthNet2021: A novel large-scale dataset and challenge for forecasting localized climate impacts

__________________________________

Cool companies I have come across this week

Health

Curai Health - A platform that uses artificial intelligence/machine learning to scale the delivery of instant medical expertise that’s accurate, trustworthy, relevant and actionable.

Climate

Climate Companion - Aiming for annual greenhouse gas reductions of more than 0.5Gt by using AI.

__________________________________

AI/ML must knows

• AutoML - The process of automating the process of applying machine learning to real-world problems. AutoML covers the complete pipeline from the raw dataset to the deployable machine learning model.
• Statsmodels - Python package for ML and statistics
• Tensorflow/keras/pytorch - Widely used machine learning frameworks
• Generative adversarial network - Generative models that create new data instances that resemble your training data. They can be used to generate fake images.
• Deep Learning - Deep learning is a form of machine learning based on artificial neural networks.

Thanks for reading and I'll see you next week!

If you are enjoying this content and would like to support the work then you can get a plan here from ￡1/month!

___________________________________

Marcel Hedman

Nural Research Founder

www.nural.cc

Marcel Hedman, is the current Choate Memorial Fellow at Harvard University focusing his efforts in data science. He also serves on the senior leadership team of the Future Leaders Network.

This newsletter is an extension of the work done by Nural Research, a group which explores AI use to inspire collaboration between those researching AI/ ML algorithms and those implementing them. Check out the website for more information www.nural.cc

Feel free to send comments, feedback and most importantly things you would like to see as part of this newsletter by getting in touch here.