Stochastic optimisation algorithms are pivotal in machine learning, especially when dealing with large datasets. These algorithms work by processing only a subset of the available data at each step, which significantly reduces computational costs while still making substantial progress towards the solution. Over the past decade, there has been a surge in research focused on leveraging stochastic optimisation for solving large-scale inverse problems. These problems often involve variational regularisation, where the solution is modelled as the minimisation of an objective function. Stochastic optimisation in this context involves various techniques such as variance reduction, acceleration, and higher-order methods. These methods are particularly useful in inverse imaging problems, where traditional machine learning approaches may not be applicable. The survey concludes with examples from imaging on linear inverse problems, highlighting the advantages and challenges of these new algorithms.
Stochastic optimisation, variance reduction, acceleration, higher-order methods
Not specified
Not specified
Not specified
Not specified
No
No
Efficient handling of large-scale data, variance reduction, acceleration techniques
No
Not specified
Not specified
Not specified
Not specified
Not specified
Not specified
No
Not specified
Not specified
Not specified
Not specified
Not specified
Not specified
Not specified
Not specified
Not specified
Inverse imaging problems
Not specified
Not specified
Not specified
Not specified
Not specified
Not specified
No
Not specified
Not specified
No
Not specified
Not specified
Not specified
Not specified
Not specified
No
Not specified
Not specified
0.00
Not specified
Not specified
01/01/1970
01/01/1970
Not specified
Not specified
Yes