View profile

AI Scholar Weekly - Issue #42

AI Scholar Weekly - Issue #42
By Educate AI  • Issue #42 • View online
AI Researchers Achieve 6.7ms on Mobile with over 78% ImageNet Accuracy; Nature-Inspired Python Optimizer; NetHack Learning Environment for RL; New AI services for DevOps and business intelligence applications (and more)

AI Success Stories
Mercy Transforms Operational Efficiency in 15 Emergency Departments (EDs) Across the System
Goal: Drive performance in Emergency Departments( EDs) across the system
The Challenge: With a new Electronic Medical Records (EMR) in place, Mercy one of the largest health systems in the US, frontline teams were drowning in data. Teams spent too much time fire-fighting and not getting ahead of operational bottlenecks. Emergency Department (ED) leaders knew that better coordination across the house was required to improve patient flow.
To achieve that, Mercy’s ED leaders needed a platform to strengthen the reliability of processes and adapt in the moment. And they needed to reduce the cognitive burden of frontline teams.
Results: Mercy adopted an AI-based solution and achieved the following results; 
  • 10% reduction in discharge Length of Stay (LOS)
  • 47% reduction in LBTC
  • >$3.5M savings from unnecessary lab utilization
  • 4% reduction in admitted LOS 
If you have a similar business need or want similar success, Sign up for a 30-minute consultation now and learn how AI can upgrade your business today. 
Top AI Research This Week
 #1  Researchers Achieve 6.7ms on Mobile with over 78% ImageNet Accuracy
With the increasing demand to efficiently deploy DNNs on mobile edge devices, it becomes much more critical to reduce unnecessary computation and increase the execution speed. Conventional methods to achieve this, including model compression and network architecture search (NAS), are primarily performed independently and do not fully consider compiler-level optimizations, a must-do for mobile acceleration. 
To address the challenge, a group of researchers has proposed 
  • A general category of fine-grained structured pruning applicable to various DNN layers
  • A comprehensive, compiler automatic code generation framework supporting different DNNs and different pruning schemes bridging the gap of model compression and NAS. 
Further, they propose NPAS, a compiler-aware unified network pruning, and architecture search. Also, to deal with large search space, they present a meta-modeling procedure based on reinforcement learning with fast evaluation and Bayesian optimization, ensuring the total number of training epochs comparable with representative NAS frameworks.
Proposed framework achieves 6.7ms, 5.9ms, 3.9ms ImageNet inference times with 78.2%, 75% (MobileNet-V3 level), and 71% (MobileNet-V2 level) Top-1 accuracy respectively on an off-the-shelf mobile phone, consistently outperforming previous work.
#2  Data Augmentation for Graph Neural Network
Data augmentation has been widely used to improve generalizability of machine learning models. However, data augmentation for facilitating GNN training has unique challenges due to graph irregularity. 
In this work, researchers tackle this problem by utilizing neural edge predictors as a means of exposing GNNs to likely (but nonexistent) edges and limiting exposure to unlikely (but existent) ones. They show that such edge predictors can encode class-homophily to promote intra-class edges and inter-class edges. 
They propose the GAUG graph data augmentation framework which uses these insights to improve node classification performance in two inference settings. Extensive experiments show the proposed GAUG-O and GAUG-M achieves up to 17% (9%) absolute F1 performance improvements across architectures and datasets, and 15% (8%) over augmentation baselines. 
#3  The NetHack Reinforcement Learning Environment
Progress in Reinforcement Learning (RL) algorithms goes hand-in-hand with the development of challenging environments that test current methods’ limits. While existing RL environments are either sufficiently complex or based on fast simulation, they are rarely both. 
This paper presents the NetHack Learning Environment (NLE), a scalable, procedurally generated, stochastic, rich, and challenging environment for RL research based on the popular single-player terminal-based roguelike game, NetHack. Researchers argue that NetHack is sufficiently complex to drive long-term research on problems such as exploration, planning, skill acquisition, and language-conditioned RL, while dramatically reducing the computational resources required to gather a large amount of experience. 
In the study, researchers compare NLE and its task suite to existing alternatives and discuss why it is an ideal medium for testing RL agents’ robustness and systematic generalization. They demonstrate empirical success for the early stages of the game using a distributed Deep RL baseline and Random Network Distillation exploration, alongside qualitative analysis of various agents trained in the environment. NLE is open source at Github
Read the paper here: The NetHack Learning Environment.
#4 Researchers Release the First of it Kind Benchmark for Generic Multiple Object Tracking
In this paper, researchers with Microsoft, Temple University, and Stony Brook University propose the first generic multiple object tracking (GMOT) benchmark named GMOT-40. 
By thoroughly considering major MOT factors and carefully annotating all tracking objects, GMOT-40 contains 40 sequences evenly distributed in 10 object categories. Associated with GMOT are two GMOT evaluation Protocols, one focusing on target association and the other on one-shot tracking. 
Several new baseline algorithms dedicated to one-shot GMOT are developed and evaluated together with relevant MOT trackers to provide references for future study. The evaluation shows that there is still considerable room to improve for GMOT, and further studies are desired. Overall, the researchers expect the benchmark to largely facilitate future research on GMOT, a critical yet under-explored computer vision problem.
#5 FairFace Challenge 2020 Results and Winning Methods 
This work presents the design and results of the FairFace Recognition Challenge at ECCV’2020. The challenge attracted 151 participants. Their submissions were evaluated on a reannotated version of IJBC database enriched by newly collected 12,549 public domain images. 
Participants were ranked using a new evaluation protocol where both accuracy and bias scores were considered. Top winning solutions obtained high performance in terms of accuracy (≥ 0.999 AUC-ROC) and bias scores. Common strategies by the participants were face pre-processing, homogenization of data distributions, the use of bias aware loss functions and ensemble models. The analysis of top-10 teams shows higher false positive rates (and lower false negative rates) for females with dark skin tone as well as the potential of eyeglasses and young age to increase the false positive rates too. See more results : FairFace Challenge at ECCV 2020
Other Great AI Papers
New Pre-Trained Image Processing Transformer that outperforms the current state-of-the-art methods. Read more 
A Nature-Inspired Python Optimizer. Read more 
The Model Counting Competition 2020 and Results. Read more 
Intrusion Detection Systems for IoT: Opportunities and Challenges offered by Edge Computing
Applying Deep-Learning-Based Computer Vision to Wireless Communications: Methodologies, Opportunities, and Challenges
AI Resources
Open source Python library focused on outlier, adversarial and drift detection. Meet Alibi Detect
AI’s most influential books for business leaders. See list 
A Python Natural Language Processing Toolkit for many human languages. Source code, documentation, and pre-trained models Available
Top AI News
Amazon launches new AI services for DevOps and business intelligence applications. Read more 
Google’s star AI ethics researcher, one of a few Black women in the field, says she was fired for a critical email
Trump’s  executive order looks to expand AI talent pipeline into agencies. Read full story
AI has cracked a problem that stumped biologists for 50 years. Read more 
This  AI-powered Santa can deliver a personalized video message to your friends or family that looks totally real. Try it Now 
AI Scholar Weekly
If this newsletter lights you up, create a ripple effect by sharing it with someone else, and they will also be lit up! And, if you have suggestions, comments, or other thoughts, we would love to hear from you, email me at, tweet at  @cdossman, like us on Facebook, or connect with me on LinkedIn
Did you enjoy this issue?
Educate AI

AI Scholar Weekly brings you everything new and exciting in the world of Artificial Intelligence, Machine Learning, and Deep Learning every week for free.

In order to unsubscribe, click here.
If you were forwarded this newsletter and you like it, you can subscribe here.
Powered by Revue