Quoting Wikipedia
Inverse problems are often ill-posed. For example, the inverse heat equation, deducing a previous distribution of temperature from final data, is not well-posed in that the solution is highly sensitive to changes in the final data.
In pattern recognition (aka machine learning), the solutions would also depend on the choice of algorithm, hyperparameters, or even random seeds used for training.
Moreover, with pattern recognition it is not even guaranteed that the solution exists, e.g. if you are conducting cluster analysis, or classification given some data, there is no guarantee that this data is suitable for such analysis. Surely, the algorithms will try their best with finding some pattern given any data, but that doesn't mean that there was any real pattern.