Profile cover photo
Profile photo
Wayne Radinsky
18,580 followers -
Software Design Engineer
Software Design Engineer

18,580 followers
About
Posts

Post has attachment
Driverless ice cream truck selling Baskin Robbins at the Hydroaviasalon science festival in Moscow.
Add a comment...

Post has attachment
nuScenes is a public large-scale dataset for autonomous driving. It enables researchers to study challenging urban driving situations using the full sensor suite of a real self-driving car. Full sensor suite (1x LIDAR, 5x RADAR, 6x camera, IMU, GPS), 1000 scenes of 20s each, 1,440,000 camera images, 400,000 lidar sweeps, two diverse cities: Boston and Singapore, left versus right hand traffic, detailed map information, manual annotations for 25 object classes, 1.1M 3D bounding boxes annotated at 2Hz, and Attributes such as visibility, activity and pose.
Add a comment...

Post has attachment
The DelFly Nimble is an agile fruit-fly-inspired robot. It's design is inspired by the wing movements of fruit flies and it can perform fruit-fly-inspired rapid banked turns.
Add a comment...

Post has attachment
"Google's launch of its latest mobile operating system, Android Pie, involves DeepMind's largest real-world machine learning roll-out to date. And there's an ambitious aim for its AI. It's looking to solve one of the modern smartphone's most frustrating features: poor battery life."

"The result was the introduction of two AI systems within the operating system. Adaptive Battery, which aims to stop apps sucking-up battery life in the background, and Adaptive Brightness, that automatically adjusts the screen depending on the environment the phone is in."

"Apps running in the background of Android devices are waking up the central processing unit (CPU) 30 per cent less and the data being transferred through Wi-Fi and mobile signals has been reduced up to 20 per cent in some cases, he says."

"We have seen a reduction in what we call the variance. Most users are familiar with the day every now and then where your battery seems to drain much faster than you're expecting or it normally does. We call these unpredictable events 'bad battery days'. We're reigning in those bad battery days.'"
WIRED UK
WIRED UK
wired.co.uk
Add a comment...

Post has attachment
TensorFlow 2.0 is designed to make TensorFlow competitive with PyTorch. TensorFlow 2.0 will use eager execution mode by default. This will simplify development. Eager mode is much easier to debug. However, graph mode will still have higher performance, the XLA (Accelerated Linear Algebra) compiler will work with graph mode, and graph mode will run on more devices. The two modes can be mixed. The new Autograph module will enable you to develop code that can run in eager mode or graph mode.

Keras will be integrated into TensorFlow to make it "Pythonic" like PyTorch. Finally, TensorFlow will eliminate deprecated APIs to make the interface clean like PyTorch.
Add a comment...

Post has shared content
The pieces were commissioned as part of an experiment by GumGum, a Santa Monica-based artificial intelligence company, and the Art & Artificial Intelligence Laboratory at Rutgers University. The researchers were not attempting to answer whether a machine could approximate a human’s final product, but rather whether a machine could approximate a human’s creative process.
Add a comment...

Post has attachment
Food recognition. "Check out the classification of recipes based off more than 400,000 food images from social media."

"In this tutorial I will show how to train deep convolutional neural networks with Keras to classify images into food categories and to output a matching recipe. The dataset contains >400'000 food images and >300'000 recipes from chefkoch.de."

"There are large overlaps in food dishes, as dishes of different categories may look very similar only in terms of image information."
Add a comment...

Post has attachment
Emo2Vec represents emotions by vectors. "We train Emo2Vec by multi-task learning six different emotion-related tasks, including emotion/sentiment analysis, sarcasm classification, stress detection, abusive language classification, insult detection, and personality recognition. Our evaluation of Emo2Vec shows that it outperforms existing affect-related representations, such as Sentiment-Specific Word Embedding and DeepMoji embeddings with much smaller training corpora. When concatenated with GloVe, Emo2Vec achieves competitive performances to state-of-the-art results on several tasks using a simple logistic regression classifier."
Add a comment...

Post has attachment
"This open challenge, which will award up to $750,000 to a range of projects, is seeking fresh and experimental approaches to four specific problems at the intersection of AI and the news."

The four problems are governing the platforms that use AI to distribute news, stopping bad actors, such as believable bots on social media to fabricating realistic video and audio, empowering journalism, by which they mean help journalists communicate about AI itself, and reimagining AI and news, where participants are encouraged to reimagine the whole AI-news ecosystem in part or entirely to better serve the public good.
AI Initiative
AI Initiative
aiethicsinitiative.org
Add a comment...

Post has attachment
"In the Atari game Pong, the agent receives a reward of either -1, 0, or +1 per step. In contrast, an agent playing Ms. Pac-Man can obtain hundreds or thousands of points in a single step. Even if the size of individual rewards is comparable, the frequency of rewards can change over time as the agent gets better. This means agents tend to focus on those tasks which have large scores, leading to better performance on certain tasks, and far worse on others."

"To resolve these kinds of issues, we developed PopArt, a technique that can adapt the scale of scores in each game so the agent judges the games to be of equal learning value, no matter the scale of rewards available in each specific game. We applied a PopArt normalisation to a state-of-the-art reinforcement learning agent, resulting in a single agent that can play a whole set of 57 diverse Atari video games, with above-human median performance across the set."
Add a comment...
Wait while more posts are being loaded