Understanding the chemical environment of pristine interfaces is a long-sought goal in electrochemistry, materials science and surface science. A substantial understanding of one such interface, the solid electrolyte interphase (SEI) in lithium anodes, originates from X-ray photoelectron spectroscopy (XPS)1,2. However, room temperature (RT) combined with ultrahigh vacuum (UHV) can induce major SEI evolution from reactions and volatilization during XPS1,2. Thus, a technique is necessary for SEI stabilization. Here we develop cryogenic (cryo)-XPS with immediate plunge freezing and demonstrate SEI preservation. We discover substantially different SEI speciation and a thicker pristine SEI with cryo-XPS, free from RT-associated thickness reduction and alterations to important species, including LiF and Li2O, in UHV. This new access to pristine SEI composition enables performance correlations across diverse electrolyte chemistries. Primarily, we highlight the necessity of studying sensitive interfaces under cryogenic conditions.

Achieving superpolynomial speed-ups for optimization has long been a central goal for quantum algorithms1. Here we introduce decoded quantum interferometry (DQI), a quantum algorithm that uses the quantum Fourier transform to reduce optimization problems to decoding problems. When approximating optimal polynomial fits over finite fields, DQI achieves a superpolynomial speed-up over known classical algorithms. The speed-up arises because the algebraic structure of the problem is reflected in the decoding problem, which can be solved efficiently. We then investigate whether this approach can achieve a speed-up for optimization problems that lack an algebraic structure but have sparse clauses. These problems reduce to decoding low-density parity-check codes, for which powerful decoders are known2,3. To test this, we construct a max-XORSAT instance for which DQI finds an approximate optimum substantially faster than general-purpose classical heuristics, such as simulated annealing. Although a tailored classical solver can outperform DQI on this instance, our results establish that combining quantum Fourier transforms with powerful decoding primitives provides a promising new path towards quantum speed-ups for hard optimization problems.

Humans and other animals use powerful reinforcement learning (RL) mechanisms that have been discovered by evolution over many generations of trial and error. By contrast, artificial agents typically learn using hand-crafted learning rules. Despite decades of interest, the goal of autonomously discovering powerful RL algorithms has proven elusive7-12. In this work, we show that it is possible for machines to discover a state-of-the-art RL rule that outperforms manually-designed rules. This was achieved by meta-learning from the cumulative experiences of a population of agents across a large number of complex environments. Specifically, our method discovers the RL rule by which the agent's policy and predictions are updated. In our large-scale experiments, the discovered rule surpassed all existing rules on the well-established Atari benchmark and outperformed a number of state-of-the-art RL algorithms on challenging benchmarks that it had not seen during discovery. Our findings suggest that the RL algorithms required for advanced artificial intelligence may soon be automatically discovered from the experiences of agents, rather than manually designed.