Large language models (LLMs) are increasingly considered as tutoring aids in science education. Yet their readiness for unsupervised use in undergraduate instruction remains uncertain, as reliable teaching requires more than fluent recall: it demands consistent, principle-grounded reasoning. Thermodynamics, with its compact laws and subtle distinctions between state and path functions, reversibility, and entropy, provides an ideal testbed for evaluating such capabilities. Here we present UTQA, a 50-item undergraduate thermodynamics question answering benchmark, covering ideal-gas processes, reversibility, and diagram interpretation. No leading 2025-era model exceeded our 95\% competence threshold: the best LLMs achieved 82\% accuracy, with text-only items performing better than image reasoning tasks, which often fell to chance levels. Prompt phrasing and syntactic complexity showed modest to little correlation with performance. The gap concentrates in finite-rate/irreversible scenarios and in binding visual features to thermodynamic meaning, indicating that current LLMs are not yet suitable for unsupervised tutoring in this domain.

The use of AI in public spaces continually raises concerns about privacy and the protection of sensitive data. An example is the deployment of detection and recognition methods on humans, where images are provided by surveillance cameras. This results in the acquisition of great amounts of sensitive data, since the capture and transmission of images taken by such cameras happens unaltered, for them to be received by a server on the network. However, many applications do not explicitly require the identity of a given person in a scene; An anonymized representation containing information of the person's position while preserving the context of them in the scene suffices. We show how using a customized loss function on region of interests (ROI) can achieve sufficient anonymization such that human faces become unrecognizable while persons are kept detectable, by training an end-to-end optimized autoencoder for learned image compression that utilizes the flexibility of the learned analysis and reconstruction transforms for the task of mutating parts of the compression result. This approach enables compression and anonymization in one step on the capture device, instead of transmitting sensitive, nonanonymized data over the network. Additionally, we evaluate how this anonymization impacts the average precision of pre-trained foundation models on detecting faces (MTCNN) and humans (YOLOv8) in comparison to non-ANN based methods, while considering compression rate and latency.

The optical properties of molecular crystals are largely determined by the excitonic coupling of neighboring molecules. This coupling is extremely sensitive to the arrangement of adjacent molecular units, as their electronic interaction is defined by the relative orientation of the individual transition dipole moments and their wave function overlap. Hence, the optical properties, such as fluorescence, are usually highly anisotropic and good indicators of structural changes during the variation of intensive thermodynamic parameters like temperature or pressure. Here, we discuss the peculiar though archetypical case of $\beta$-phase zinc-phthalocyanine: In single crystals, we report a sudden change of spectral emission with temperature from a broad, unpolarized Frenkel-exciton type luminescence to a narrow, highly polarized superradiance-like fluorescence below 80 K. Surprisingly, we find that there is no sign of a discrete structural phase transition in this temperature regime. To understand this apparent contradiction, we perform polarization-, temperature- and time-dependent photoluminescence measurements along different crystallographic directions to fully map the emission characteristics of the crystal-exciton. By means of ab-initio calculations on a density functional theory level we conclude that our observations are consistent with a dimer exciton model when considering thermalized electronic states. As such, our study presents a representative case study on a well-established molecular material class demonstrating that caution is advised when attributing discrete changes in electronic observables to a structural phase transition. As we show for zinc-phthalocyanine in its $\beta$-phase modification, slowly varying excitonic couplings and thermal redistribution of excitations can mimic the same signatures attributed to a structural phase transition.