• AI Research Insights
  • Posts
  • AI News: Deepmind's Sparrow Beta Coming!; Stanford's Emotion Generator;Interesting Conclusion On LLMs; UCLA's ClimateLearn Library......

AI News: Deepmind's Sparrow Beta Coming!; Stanford's Emotion Generator;Interesting Conclusion On LLMs; UCLA's ClimateLearn Library......

Hi there, today we will be sharing some research updates from Deepmind, Stanford, Interesting Conclusion On LLMs, UCLA, MIT, Shanghai AI Laboratory, CMU, AI2, and some bonus cool AI tools. So, let's start...

Deepmind: In response to OpenAI's 'ChatGPT,' DeepMind is considering a private beta release of their 'Sparrow' chatbot sometime in 2023.

Snap/Stanford: Now There's A Model That Generates Emotions And Descriptions Recalled From Real-world Images. The research proposes Affective Explanation Captioning (AEC), a task that generates emotions and explanations recalled from real-world images. 6283 annotators felt emotions and explanations for 85007 real-world images created Affection, a large dataset of annotated feelings and descriptions of 85007 real-world images by 6283 annotators. Turing test showed that about 40% of evaluators could not discriminate between a neural speaker created using Affection and a human.

Interesting Conclusion On LLMs: These two papers suggest that few-shot prompting LLMs may be more similar to fine-tuning than realized.

UCLA: Introducing ClimateLearn, a new PyTorch library for accessing climate datasets, state-of-the-art ML models, and high quality training and visualization pipelines

MIT: MIT researchers determined that 1 billion autonomous vehicles, each driving for one hour per day with a computer consuming 840 watts, would consume enough energy to generate about the same amount of emissions as data centers currently do.

Shanghai AI Laboratory/ Tsinghua University: This research from China introduces InternImage, a large-scale CNN model that outperforms ViT. Experiments on ImageNet, COCO, ADE20K, and a wide range of benchmarks demonstrated that InternImage achieves accuracy comparable to or better than ViT trained on large data sets, showing that CNNs are a viable option for large-scale models. This demonstrates that CNNs have great potential as an option for large-scale models. One challenge is that DCN-based methods are processing-heavy.

CMU: A deep neural network maps the phase and amplitude of WiFi signals to UV coordinates within 24 human regions. The model can estimate the dense pose of multiple subjects by utilizing WiFi signals as the only input.

AI2: Researchers From Allen Institute for AI Introduce TeachMe: A Framework To Understand And Correct AI Models. TeachMe is the suggested system, which consists of two key parts: (1) Entailer, a T5-based machine reasoning model that can generate valid lines of reasoning. (2) a dynamic database of previous comments.

New Unicorn: Germany’s Translation Tech Firm DeepL Becomes Latest AI Unicorn. DeepL has emerged as one of the most disruptive players in the machine translation space with a product that rivals even Google’s technology.

Cool Tool: Airtest generates unit tests using AI for C#, C++, Go, Java, JavaScript, PHP, Python, R Lang, Ruby, and Swift code. Simply paste a block of code, choose your language and testing framework, then hit "Generate" to witness magic.

How did we do?

Did you like today's newsletter?

Login or Subscribe to participate in polls.