On Friday, 5 June 2026, Quantum@SUN and NITheCS hosted Mpho Tladi from the University of Cape Town for a seminar titled “Stochastic Krylov Dynamics: Revisiting Operator Growth in Open Quantum Systems.”
The seminar explored one of the big questions in modern quantum physics: what happens to quantum information when a system is no longer isolated from its environment?
Tladi took the audience through the idea of Krylov complexity, a framework used to track how a simple quantum operator grows, spreads, and becomes more complex over time. A key moment in the talk was the explanation of the Krylov chain, where operator growth can be imagined as movement along a one-dimensional chain. The further the operator moves, the more complex it becomes.
The seminar then moved from closed quantum systems, where the evolution is cleaner and more controlled, to open quantum systems, where the environment introduces noise, decoherence, and dissipation. Using tools such as Lindblad dynamics and the Schwinger–Keldysh formalism, Tladi showed that environmental effects do not simply disturb quantum chaos — they can reshape how complexity grows.
One of the strongest takeaways was the idea that open quantum systems are caught in a race. On one side, the system tries to scramble information through its internal quantum dynamics. On the other, the environment tries to suppress or destroy that growth. If scrambling happens fast enough, the system can still behave like a closed chaotic system. But if the environment wins, operator growth is cut off before information can fully spread.
The discussion after the talk also opened up thoughtful questions around emergent phase space, non-Hermitian dynamics, symplectic structure, and the physical meaning of Lanczos coefficients in the Krylov-chain picture.
Overall, the seminar showed that open quantum systems are not just messy versions of closed systems. They bring a different kind of physics, where noise and dissipation can decide whether information scrambling survives at all.
For more on the seminar, click here:
