This project combines recent quantum information-theoretic methods from quantum communication and cryptography with ML techniques to develop entanglement measures suitable for many-body systems. For this we take an operational approach, quantifying entanglement relative to specific tasks, where entanglement can be exploited for enhanced efficiency. This operational approach also addresses computational considerations: our measures account only for entanglement that can be feasibly extracted. Ultimately, we aim to apply these novel entanglement measures to characterize entanglement strucutres in many-body states realized in present-day quantum simulators.