MIT researchers have developed neural “liquid” network that varies its equations parameters, enhancing its ability to analyze time series data.
Image: Jose-Luis Olivares, MIT
[mks_button size=”medium” title=”Exploration Highlights” style=”rounded” url=”http://” target=”_self” bg_color=”#0000FF” txt_color=”#FFFFFF” icon=”” icon_type=”” nofollow=”0″]
- The new type of neural network could aid decision making in autonomous driving and medical diagnosis.
- These flexible algorithms, dubbed “liquid” networks, change their underlying equations to continuously adapt to new data inputs. The advance could aid decision making based on data streams that change over time, including those involved in medical diagnosis and autonomous driving.
- “This is a way forward for the future of robot control, natural language processing, video processing — any form of time series data processing,” says Ramin Hasani
- Hasani designed a neural network that can adapt to the variability of real-world systems.
- This flexibility is key. Most neural networks’ behavior is fixed after the training phase, which means they’re bad at adjusting to changes in the incoming data stream. Hasani says the fluidity of his “liquid” network makes it more resilient to unexpected or noisy data, like if heavy rain obscures the view of a camera on a self-driving car. “So, it’s more robust,” he says.
- Hasani’s network excelled in a battery of tests. It edged out other state-of-the-art time series algorithms by a few percentage points in accurately predicting future values in datasets, ranging from atmospheric chemistry to traffic patterns.
- “We think this kind of network could be a key element of future intelligence systems.”
Original Article: “Liquid” machine-learning system adapts to changing conditions
More from: Massachusetts Institute of Technology on Innovation2
Go deeper with the LATEST Bing News on:
Liquid neural network
- Neural Correlates of Motor Dysfunction in Children with Traumatic Brain Injury: Exploration of Compensatory Recruitment Patternson April 10, 2021 at 4:59 pm
Visual templates, announcing the task to be performed, were projected by a Barco 6400i liquid crystal display projector (Kortrijk, Belgium, 1024 × 768 pixels, 60 Hz) and displayed via a mirror ...
- Partnership to deploy AI tech in single-use endoscopeson April 8, 2021 at 7:18 pm
NexOptic's ALIIS is the company's proprietary neural network technology, and it uses machine learning for imaging. Pristine Surgical's endoscopes will incorporate the technology for image optimization ...
- Electrical and Computer Engineeringon April 8, 2021 at 2:02 pm
Topics include truth and underdetermination, induction, computability, language learning, pattern recognition, neural networks, and the role of simplicity ... junction lasers, liquid crystal devices, ...
- Facial recognition for bears (and other ways to use the technology for good)on April 8, 2021 at 3:30 am
But that’s people. For bears, it’s all good – and facial recognition is now being used to help research, monitor and protect the animals using a neural network-based system called BearID.
- Primate ventral striatum maintains neural representations of the value of previously rewarded objects for habitual seekingon April 8, 2021 at 2:44 am
Ventral striatum is known to be involved in the value update for habit learning. Here, the authors report neural and behavioural correlates for the long-term maintenance of value memory for previously ...
- What is deep learning?on April 7, 2021 at 7:50 am
Deep learning refers to a technique for creating artificial intelligence (AI) using a layered neural network, much like a simplified replica of the human brain. It fits into a larger family of ...
- Engineered glycomaterial implants orchestrate large-scale functional repair of brain tissue chronically after severe traumatic brain injuryon April 4, 2021 at 5:01 pm
Chondroitin sulfate glycosaminoglycans (CS-GAGs) are key regulators of growth factor signaling and neural stem cell homeostasis in the brain. However, the efficacy of engineered CS (eCS) matrices in ...
- AI Systems Startup Scales Argonne’s Cosmic Cleanupon March 29, 2021 at 7:16 am
For neural networks to expand into more areas in traditional supercomputing ... In recent cosmology work, Argonne found that by modifying ResNet to optimize for cosmic background remove from liquid ...
- iPhone 11 vs. iPhone 12: Which should you buy in 2021?on March 18, 2021 at 5:01 pm
The iPhone 11 features an LCD “Liquid Retina” panel with a 1792-by ... The iPhone 12 also features the latest Neural Engine for improved machine learning features and capabilities.
- Power/Performance Bits: March 16on March 15, 2021 at 5:00 pm
In the equations used to structure the neural network, parameters are allowed to change over time based on the results of a nested set of differential equations. Hasani said that the ‘liquid’ network ...
Go deeper with the LATEST Bing News on:
- A graph neural network framework for causal inference in brain networkson April 13, 2021 at 7:27 am
A central question in neuroscience is how self-organizing dynamic interactions in the brain emerge on their relatively static structural backbone. Due to the complexity of spatial and temporal ...
- Smacking your child can affect their BRAIN development and disrupt the neural regions that affect decision-making and processing of situations, study warnson April 13, 2021 at 7:24 am
Smacking appears to alter children's neural responses to environmental threats in a manner similar to more severe forms of maltreatment, scientists reveal.
- Deeplite raises $7.5 million CAD seed round to optimize deep neural networkson April 13, 2021 at 7:18 am
Montréal-based Deeplite has raised $7.5 million CAD in seed funding as the startup looks to accelerate its software that optimizes deep neural networks.
- Perceiver: One Neural-Network Model for Multiple Input Data Typeson April 13, 2021 at 7:16 am
The great breakthrough about this model is that it makes no assumption about input data type, while, for instance, existing convolutional neural networks work for images only. Source: Perceiver: ...
- A review and comparison of breast tumor cell nuclei segmentation performances using deep convolutional neural networkson April 13, 2021 at 5:57 am
This study's objective was to determine if deep convolutional neural networks can be trained, with transfer learning, on a set of histopathological images independent of breast tissue to segment tumor ...
- Study reveals neural stem cells age rapidlyon April 12, 2021 at 9:32 am
In a new study published in Cell Stem Cell, a team led by USC Stem Cell scientist Michael Bonaguidi, Ph.D., demonstrates that neural stem cells—the stem cells of the nervous system—age rapidly.
- DeepONet: A deep neural network-based model to approximate linear and nonlinear operatorson April 12, 2021 at 6:30 am
Artificial neural networks are known to be highly efficient approximators of continuous functions, which are functions with no sudden changes in values (i.e., discontinuities, holes or jumps in graph ...
- Neural Network Market 2021 : Top countries Data, Market Size, In-Depth Analysis with Booming Trends and Segmentation Forecast to 2026on April 8, 2021 at 11:26 pm
Global “Neural Network Market" 2021-2026 Research Report provides key analysis on the market status of the Neural ...
- Deep Neural Networks Market Outlook, Demand and Supply, Key Prospects, Pricing Strategies, Forecast and Top Manufacturers Analysis Reporton April 8, 2021 at 5:19 am
The MarketWatch News Department was not involved in the creation of this content. Apr 08, 2021 (Market Insight Reports) -- The Global Deep Neural Networks Market is projected to reach USD 5.98 billion ...
- Deep learning networks prefer the human voice -- just like uson April 6, 2021 at 2:54 pm
The researchers discovered that in a side-by-side comparison, a neural network whose 'training labels' consisted of sound files reached higher levels of performance in identifying objects in images, ...