Generative AI (GenAI) has shifted AI capabilities from discriminative prediction to creative interaction, offering opportunities to augment productivity and innovation. However, realizing these benefits requires navigating risks where development outpaces governance. This article revisits the Six Human-Centered AI (HCAI) Grand Challenges to analyze their relevance in the generative era. Critical new requirements are identified: preserving human autonomy, ensuring operational safety against non-deterministic outputs, and navigating complex intellectual property landscapes. These findings are synthesized into an updated, actionable research agenda for each challenge, serving as a call to action to operationalize these principles. By shifting focus from risk mitigation to human empowerment, this agenda establishes human-centeredness as the organizing principle for a future where GenAI enhances human agency, dignity, and collective flourishing.
Understanding the structure and dynamics of ring or cyclic polymers is a long-standing challenge in polymer science, with important implications for emerging biological phenomena such as chromosome territories. This Perspective article provides a comprehensive overview of the current state of ring polymer physics and rheology, highlighting emerging challenges and opportunities for future research. Key scientific questions are considered regarding the properties of synthetic and biological ring polymer systems using theory, simulations, and experiments. This article was inspired by stimulating discussions at a CECAM Flagship workshop on Ring Polymer Dynamics in Prato, Italy, in June 2023. Several of the concepts and results discussed here are also presented in the Journal of Rheology virtual issue on ring polymers (https://pubs.aip.org/jor/collection/1392/Ring-Polymers). Broadly, this article aims to spark conceptual advances in polymer physics and rheology by exploring new phenomena and open scientific questions that are unique to ring polymer systems.
We propose an efficient protocol to realize multi-qubit gates in arrays of neutral atoms. The atoms encode qubits in the long-lived hyperfine sublevels of the ground electronic state. To realize the gate, we apply a global laser pulse to transfer the atoms to a Rydberg state with strong blockade interaction that suppresses simultaneous excitation of neighboring atoms arranged in a star-graph configuration. The number of Rydberg excitations, and thereby the parity of the resulting state, depends on the multiqubit input state. Upon changing the sign of the interaction and de-exciting the atoms with an identical laser pulse, the system acquires a geometric phase that depends only on the parity of the excited state, while the dynamical phase is completely canceled. Using single qubit rotations, this transformation can be converted to the C_kZ or C_kNOT quantum gate for k+1 atoms. We also present extensions of the scheme to implement quantum gates between distant atomic qubits connected by a quantum bus consisting of a chain of atoms.
Mucormycosis is an emerging, life-threatening human infection caused by Mucorales fungi1-3. Metabolic disorders uniquely predispose an ever-expanding group of patients to mucormycosis through poorly understood mechanisms1,2,4,5, suggesting that uncharacterized host metabolic effectors may confer protective immunity against this infection. Here we uncover a master regulatory role of albumin in host defence against Mucorales through the modulation of fungal pathogenicity. Our initial studies identified severe hypoalb uminaemia as a prominent metabolic abnormality and an independent biomarker of poor mucormycosis outcome across three distinct cohorts of patients with mucormycosis. Notably, purified albumin selectively inhibits Mucorales growth among a range of pathogens, and albumin-deficient mice display susceptibility specifically to mucormycosis. The antifungal activity of albumin is mediated by the release of bound free fatty acids (FFAs). Albumin prevents FFA oxidation, which otherwise abolishes their antifungal properties, and sera from patients with mucormycosis display high levels of oxidized FFAs. Physiologically, albumin-bound FFAs suppress the expression of key virulence factors by inhibiting protein synthesis, the reby rendering Mucorales avirulent in vivo. Overall, we identify a host defence mechanism that directs the pathogen to suppress its pathogenicity program in response to unfavourable metabolic cues regulated by albumin. These findings have major implications for the pathogenesis and management of mucormycosis.
Learning joint representations across multiple modalities remains a central challenge in multimodal machine learning. Prevailing approaches predominantly operate in pairwise settings, aligning two modalities at a time. While some recent methods aim to capture higher-order interactions among multiple modalities, they often overlook or insufficiently preserve pairwise relationships, limiting their effectiveness on single-modality tasks. In this work, we introduce Contrastive Fusion (ConFu), a framework that jointly embeds both individual modalities and their fused combinations into a unified representation space, where modalities and their fused counterparts are aligned. ConFu extends traditional pairwise contrastive objectives with an additional fused-modality contrastive term, encouraging the joint embedding of modality pairs with a third modality. This formulation enables ConFu to capture higher-order dependencies, such as XOR-like relationships, that cannot be recovered through pairwise alignment alone, while still maintaining strong pairwise correspondence. We evaluate ConFu on synthetic and real-world multimodal benchmarks, assessing its ability to exploit cross-modal complementarity, capture higher-order dependencies, and scale with increasing multimodal complexity. Across these settings, ConFu demonstrates competitive performance on retrieval and classification tasks, while supporting unified one-to-one and two-to-one retrieval within a single contrastive framework.
