Menu
Your Data and
AI Journey
Does your organization sell ocean-related products or services?
Do you collect store and use a lot of quality data?
Do you have strong technology/IT skills internally?
Do you have plans and processes in place today to guide how your organization wants to leverage data?
Have you considered expanding into the ocean ecosystem?
Did you know, the worldwide ocean economy is valued at around US$1.5 trillion per year. 80-90% of global trade by volume is carried by sea. 350 million jobs world-wide are linked to fisheries.
We would be happy to make some intros!
Connect with us to find out how DeepSense can make AI and ML possible for you:
Awesome. Let's explore a possible AI/machine learning solution for you to use right away.
Perfect. We can help support your tech team.
Let's explore a prototype design to help guide your future AI/ML opportunities.
Machine LearningMachine Learning
DeepSense is designed to help companies leverage machine learning. Machine learning is a computer program developed to automatically extract patterns to optimize a performance based on past experience or example data. Essentially, computers learn to make a prediction without being specifically programmed. Example: Determine if a pipe has corrosion, the sentiment of text in a document, or the estimated delivery time of a shipment.
Machine LearningSupervised LearningSupervised Learning
A machine learning method that determines a predictive model usingdata points with known outcomes. With labeled data, a model is trained using feedback to optimize a prediction based on input data. Commonly used supervised learning algorithms are decision trees, logistic regression, and support vector machine.Example: Predict product sale based on past sales, a chatbot, or predict a temperature value from historical values.A machine learning method that determines a predictive model using data points with known outcomes. With labeled data, a model is trained using feedback to optimize a prediction based on input data. Commonly used supervised learning algorithms are decision trees, logistic regression, and support vector machine. Example: Predict product sale based on past sales, a chatbot, or predict a temperature value from historical values.
Machine LearningSupervised LearningClassification
A machine learning method that determines a predictive model usingdata points with known outcomes. With labeled data, a model is trained using feedback to optimize a prediction based on input data. Commonly used supervised learning algorithms are decision trees, logistic regression, and support vector machine.Example: Predict product sale based on past sales, a chatbot, or predict a temperature value from historical values.A form of supervised learning, classification algorithms are focused on identifying in which predetermined category data belongs. Example: Predict if equipment will fail: Yes/No or classify if a product falls into one of multiple categories: Good/Bad/Unsure.
Machine LearningSupervised LearningLinear Regression
One of the most common supervised learning algorithms used, a regression provides a simple model to make a numeric value prediction. Use case: Predict the price or value water temperature A regression problem is when the output variable is a real value, such as “dollars” or “weight”.
Machine Learning
Neural Networks
Neural networks attempt to emulate the neurological pathways of the human brain. There will always be at least three layers in a neural network: an input layer, a hidden layer and an output layer. There can be more than one hidden layer. Neural networks handle extremely complex tasks such as image recognition. Neural networks are slow to train, requiring heavy computation requirements and are very hard to understand (not explainable).Example: Mammal recognition using video footage or images.
Machine Learning
Unsupervised Machine Learning
The goal for unsupervised learning is to model the underlying structure or distribution in the data in order to learn more about the data. Unlike supervised learning any data labels are not used to train the algorithm. Algorithms work to discover patterns in the data. Common unsupervised learning algorithms are k-means clustering, hierarchical clustering, and apriori algorithm.
Machine Learning
Data Sets
Data is typically split into training, test and validation sets. A validation set can be used to test the performance when more than one algorithm is developed. Without appropriate data segmentation there is risk the model does not fully learn from all data
Fantastic. Let's plan a strategic workshop to map future possibilities.
Connect with us to find out how DeepSense can make AI and ML possible for you:
