Alibaba Open-Sources Its MCU to Boost AI Research
Alibaba’s chip subsidiary Pingtouge (平头哥) has become the first Chinese company to open-source its Microcontroller (MCU) design platform.
Author: Synced
Machine Intelligence | Technology & Industry | Information & Analysis
Harvard & Google Seismic Paper Hit With Rebuttals: Is Deep Learning Suited to Aftershock Prediction?
With deep learning emerging as something of a panacea in the world of science, AI researchers and seismologists alike are leveraging the tech in pursuit of better aftershock forecast solutions.
Can Pretrained Language Models Replace Knowledge Bases?
The recent rapid development of pretrained language models has produced significant performance improvements on downstream NLP tasks.
Robots to Handle Rubik’s Cube, Push Stuff Around, Build a Lunar Base
Synced Global AI Weekly October 20th
AI In Wildlife Conservation
Image recognition technology in particular has come to play a valuable role in wildlife conservation, where endangered species are tallied and tracked so their numbers and migrations can be more accurately measured and understood.
Smooth Exclusion: New Adobe Algorithm Aces Video Inpainting
Purging your favourite photos or videos of an unsightly trash pile, a parked car or even an ex-partner has never been easier, thanks to the rapid progress of AI models designed for such tasks.
Kaldi Creator Daniel Povey Joining Xiaomi in Beijing
Daniel Povey, the main developer of the widely used open-source speech recognition toolkit Kaldi, tweeted today that he is likely joining Chinese smartphone giant Xiaomi at its Beijing headquarters to work on a next generation “PyTorch-y Kaldi.”
Sotabench: Benchmarking Open Source Models Directly From GitHub
Machine learning research resource Papers with Code last week introduced Sotabench, a free and open website created to benchmark and rate the performance of state-of-the-art open source models from GitHub.
Using Conditional GANs to Build Zelda Game Levels
A team of researchers has proposed a Generative Adversarial Network based model tasked with creating playable and aesthetically appealing game levels for popular action-adventure video game series The Legend of Zelda.
OpenAI Robot Hand: Today Rubik’s Cube, Tomorrow the Real World?
We’ve lost the brain race, but humans still have unmatched dexterity, right? Wrong. OpenAI’s humanlike five-fingered gripper Dactyl just single-handedly solved a Rubik’s cube.
Shake Your Booty: AI Deepfakes Dance Moves From a Single Picture
Do you have two left feet? Do you avoid the dance floor out of fear of embarrassment? If you’ve ever secretly wished you could move your body like Joaquín Cortés — well, at least in a video — a new AI-powered 3D body mesh recovery module called Liquid Warping GAN can give you a leg up.
PyTorch 1.3 Released; News From Arm TechCon 2019; Are Deepfakes More Dangerous Than Imagined?
Synced Global AI Weekly October 13rd
How AI Is Modernizing Chinese Agriculture
In recent years frontier technologies such as AI, big data, IoT, and 3S technology (RS remote sensing technology, GIS geographic information systems, and GPS global positioning systems), have been aggressively deployed to accelerate the modernization of Chinese agriculture.
OpenCV-Inspired Kornia Is a Differentiable Computer Vision Library for PyTorch
Inspired by OpenCV, Kornia is based on PyTorch and designed to solve generic computer vision problems.
Facebook Debuts PyTorch 1.3 With PyTorch Mobile, Quantization, TPU Support and More
The latest version, PyTorch 1.3, includes PyTorch Mobile, quantization, and Google Cloud TPU support. The release was announced today at the PyTorch Developer Conference in San Francisco.
Open Source Hardware Trends, Arm Takes a Different Tack
Although the term “open source” conveys sentiments such as research sharing and community building, leading semiconductor IP provider Arm, which supports 95 percent of smartphone embedded processors, is not a fan.
Is NVIDIA Doubling Down On RISC-V?
The US chip giant now appears to be exploring the terrain beyond GPU architecture, raising the stakes on other dedicated electronic circuits.
Watch Out, MIT’s New AI Model Knows What You’re Doing Behind That Wall
For better or worse, AI can now figure out what you’re doing even without “seeing” you. The MIT Computer Science & AI Lab (CSAIL) has unveiled a neural network model that can detect human actions through walls or in extremely dark places.
ChipGAN Style Transfer Masters Chinese Ink Wash Painting
In the paper ChipGAN: A Generative Adversarial Network for Chinese Ink Wash Painting Style Transfer, a team of researchers from Peking University and Tsinghua University propose an end-to-end GAN-based architecture that can transfer input photos into the style of Chinese ink wash paintings.
Huawei’s TinyBERT Is 7X Smaller and 9X Faster Than BERT
Researchers from the Huazhong University of Science and Technology and Huawei Noah’s Ark Lab have introduced TinyBERT, a smaller and faster version of Google’s popular large-scale pre-trained language processing model BERT.
TensorFlow 2.0 Released; Google’s Large Scale Multilingual Speech Recognition; AI Responds to Climate Change
Synced Global AI Weekly October 6th
Can AI Assist in Suicide Prevention?
Figures from nonprofit organization Mental Health America show the severe depression rate for US youth increased from 5.9 to 8.2 percent over the last five years. The numbers provide a sobering glimpse of the shortcomings in mental health care, and some researchers are suggesting AI might be able to address the problem.
Google Accelerates Quantum Computation with Classical Machine Learning
Tech giant Google’s recent claim regarding quantum supremacy created a buzz in the computer science community and got global mainstream media talking about quantum computing breakthroughs.
Hugging Face Implements SOTA Transformer Architectures for PyTorch and TensorFlow 2.0
NLP-focused startup Hugging Face recently released a major update to their popular “PyTorch Transformers” library which establishes compatibility between PyTorch and TensorFlow 2.0.
NVIDIA & ORNL Researchers Train AI Model on World’s Top Supercomputer Using 27,600 NVIDIA GPUs
A research team from NVIDIA, Oak Ridge National Laboratory (ORNL), and Uber has introduced new techniques that enabled them to train a fully convolutional neural network on the world’s fastest supercomputer, Summit, with up to 27,600 NVIDIA GPUs.
DeepMind Uses GANs to Convert Text to Speech
DeepMind researchers are expanding GANs to audio, with a new adversarial network approach for high fidelity speech synthesis.
‘AI in Action’ | Toronto Hosts AI Squared Forum 2019
On September 22 about 300 artificial intelligence (AI) and deep learning enthusiasts gathered at the University of Toronto Bahen Centre for Information Technology for the AI Squared Forum 2019.
GitHub Releases Dataset of Six Million Open-Source Methods for Code Search Research
In a bid to help software developers and foster innovative code search research, GitHub last week announced the CodeSearchNet Challenge in a joint effort with California-based machine learning development tools startup Weights & Biases.
What Does ‘Broken’ Sound Like? First-Ever Audio Dataset of Malfunctioning Industrial Machines
An R&D team from Japanese multinational conglomerate Hitachi has proposed a similar, sound-based technique for identifying malfunctioning industrial robots on a factory floor.
The Wait is Over -TensorFlow 2.0 Released!
Google today announced the final release of TensorFlow 2.0. This follows the platform’s alpha release at the TensorFlow Dev Summit in June.
Picture Perfect Beauty Courtesy AI Makeup Artist
A multi-institute research group recently released the paper PSGAN: Pose-Robust Spatial-Aware GAN for Customizable Makeup Transfer, which proposes a novel method for transferring makeup styles from a reference picture to a user’s source image.
Boston Dynamics Lets the Dogs Out; Google Releases Deepfake Detection Dataset
Synced Global AI Weekly September 29th
Free (Robot) Hugs! An Embracing Multimodal Dataset
In an attempt to unlock the science behind hugs, a team of researchers from Arizona State University slapped wearable sensors on 33 humans and collected data on their more than 350 huggings with the humanoid remote-controlled robot “Baxter.”
Google’s ALBERT Is a Leaner BERT; Achieves SOTA on 3 NLP Benchmarks
Google’s new “ALBERT” language model has achieved state-of-the-art results on three popular benchmark tests for natural language understanding (NLU): GLUE, RACE, and SQuAD 2.0.
ARM and Qualcomm Will Continue to Supply Huawei Despite US Ban
Chinese news outlet Caixin is reporting that UK-based chip designer ARM confirmed on Wednesday it will continue to license its chip design architecture to Chinese tech giant Huawei and the company’s chip unit HiSilicon Technologies despite a US trade ban.
Training SOTA Language Models With Three Lines of Code
Hugging Face, a startup specializing in natural language processing, today released a landmark update to their popular Transformers library, offering unprecedented compatibility between two major deep learning frameworks, PyTorch and TensorFlow 2.0.
Reproducibility Challenges in Machine Learning for Health
A team of researchers have proposed a number of “recommendations to data providers, academic publishers, and the ML4H research community in order to promote reproducible research moving forward” in their new paper Reproducibility in Machine Learning for Health.
What’s Cooking? Google VideoBERT Predicts Recipes
This summer Google Research extended BERT to video and introduced VideoBERT, a self-supervised learning model that can learn the relationship between videos and texts.
Alibaba’s New AI Chip Can Process Nearly 80K Images Per Second
At the Alibaba Cloud (Aliyun) Apsara Conference 2019, Pingtouge unveiled its first AI dedicated processor for cloud-based large-scale AI inferencing. The Hanguang 800 is the first semiconductor product in Alibaba’s 20-year history.
Researchers Use Vile Comments from Trump Subreddit to Train AI to Battle Hate Speech
Researchers from Intel AI and University of California at Santa Barbara have introduced a new generative hate speech intervention model, along with two large-scale fully-labeled hate speech datasets collected from Reddit and Gab.