Quantum computing presents computational challenges ranging from the simulation of large quantum circuits to the inference of errors and the search for optimal solutions. Although these problems involve enormous spaces of possibilities, their structure can offer opportunities for more efficient computation. This talk presents recent advances in addressing these challenges using tensor networks, statistical mechanics, and machine learning. I will discuss our work on classical simulation of quantum circuits, including benchmarks of quantum computational advantage and the validation of experimental outputs from Google’s Sycamore circuits and IBM’s doped-Clifford circuits. I will then explore how statistical-mechanical formulations, exact decoding, and learned noise models can improve quantum error correction. Extending these ideas to quantum-inspired optimization, I will introduce our physics-inspired Free Energy Machine framework and its application to combinatorial problems. Finally, we will discuss how these computational methods can inform the design of quantum experiments and improve the reliability of logical qubits.

Semiconductors were first measured in the 1850s, and their counter-intuitive properties confused & frustrated scientists for nearly a century. The semiconductor transistor was only invented in 1947, making semiconductors useful for computers. With the discovery of graphene in 2004, scientists became fascinated by semimetals. Our community worldwide has now explored the compensated semimetal WTe₂, the Dirac semimetal Na₃Bi, the polar Weyl semimetal TaAs, the Weyl ring semimetal Co₂MnGa and a zoo of other materials [1-6]. Despite their fascinating physics, semimetals are not yet useful. I argue that driving materials across the phase transition—between semimetal and semiconductor—is the most natural way to operate a semimetal. We recently turned a Weyl semimetal ‘on’ and ‘off’ for the first time in (Cr,Bi)₂Te₃, using chemical doping & temperature to drive the phase transition to a semiconductor [7-8]. Now, we need to find more materials exhibiting a semimetal-semiconductor transition, drive the transition at high frequency [9], and understand how such ‘semimetal switches’ could surpass semiconductor transistors.

As the dominant mode of interannual climate variability, El Niño–Southern Oscillation (ENSO) significantly influences weather, marine and terrestrial ecosystems, and human society across the globe, making reliable prediction critically important. Predicting ENSO a few seasons in advance remains one of the major challenges in climate science, particularly across the Spring Predictability Barrier and for prolonged or atypical ENSO events. This presentation examines ENSO predictability within the broader context of Pacific climate variability, with an emphasis on the physical processes that govern ENSO evolution and provide sources of predictability. In particular, we explore the interaction among tropical ocean memory, ocean–atmosphere coupling, and extratropical forcing using a physically based statistical ENSO Prediction Model (EPM). The contrasting 1998–2001 and 2020–2023 triple-dip La Niña events illustrate how similar ENSO states can arise from different physical mechanisms: the former was largely sustained by negative equatorial Pacific heat-content anomalies following a strong El Niño, whereas the latter was strongly influenced by persistent extratropical forcing from the Southern Hemisphere. The EPM captures these distinct processes, demonstrating how physical understanding of tropical–extratropical interactions can improve ENSO prediction. Building on these insights, we examine the ongoing development of the 2026/27 El Niño to explore how evolving ocean heat content, atmospheric forcing, and tropical–extratropical interactions may provide early indications of its future evolution and test our current understanding of ENSO predictability.

Engineering light at the nanoscale has opened exciting opportunities for exploring fundamental light–matter interactions and developing next-generation photonic and quantum technologies. By tailoring the geometry, composition, and electromagnetic resonances of nanostructures, it is able to achieve unprecedented control over optical confinement, scattering, absorption, and emission, far beyond the capabilities of conventional optical components. In this plenary talk, I will present our recent advances in nanoscale light–matter interactions, spanning classical nanophotonics, nonlinear optics, and quantum photonic devices. First, I will discuss how advanced sub-10-nm nanofabrication enables precisely engineered dielectric and plasmonic nanostructures for strong optical confinement, tailored resonances, and enhanced light absorption, with applications in multifunctional optoelectronic devices. Next, I will highlight how Fano resonances and bound states in the continuum provide powerful platforms for enhancing nonlinear optical processes and manipulating light emission. Finally, I will introduce our recent efforts in quantum nanophotonics, where engineered optical environments enable enhanced light–emitter coupling, controlled spontaneous emission, and tailored single-photon emission. Together, these developments demonstrate how nanoscale engineering bridges classical optical phenomena and quantum light–matter interactions, opening new opportunities for integrated photonics and emerging quantum technologies.

Unconventional antiferromagnets (AFs), such as noncollinear AFs (NCAF) and altermagnets, have garnered significant interest due to their compensated spin structures and spin-splitting properties which enable functionalities beyond those of conventional ferromagnets (FMs) and AFs [1]. Notably, unconventional AFs combine the advantages of both FMs and AFs – offering field immunity, ultrafast spin dynamics, and energy-efficient operation – making them promising candidates for next-generation computing and sensing technologies.

First, we discuss NCAF material platforms based on hexagonal-close-packed (002)-oriented Mn₃Ga and face-centered-cubic (111)- and (001)-oriented Mn3Pt thin films grown using magnetron sputtering [2-4]. In Mn₃Ga/Pt heterostructures, ultrafast spin current generation and THz signal arise from pulsed magnetization dynamics and transient spin currents associated with NCAF order [3]. In Mn₃Pt multilayers, modulation of the lattice parameter and dopant concentration enables control of the interplay between ferromagnetic canting and intrinsic antiferromagnetic contributions, providing tunability of the magnitude and polarity of the anomalous Hall effect and x-ray magnetic signatures [4].

Next, we further explore the translation of field-tolerant and magnetically compensated nature of NCAFs into magnetic-field sensing technology for demanding environments such as magnetic resonance imaging, where conventional magnetic sensors can be challenged by strong background fields and magnetic interference.

Finally, we discuss the development of altermagnetic thin films, including pyrite MnS₂ and rutile RuO₂, grown using chemical and physical vapour deposition, respectively. Combining density functional theory predictions of momentum-dependent spin-splitting near the Fermi level with experimental characterization of their chemical, structural, magnetic and electrical properties, we establish these materials as promising platforms for investigating and exploiting altermagnetic functionality [5].