arXiv:2602.10056v2 Announce Type: replace
Abstract: We introduce WildCat, a high-accuracy, low-cost approach to compressing the attention mechanism in neural networks. While attention is a staple of modern network architectures, it is also notoriously expensive to deploy due to resource requirements that scale quadratically with the input sequence length $n$. WildCat avoids these quadratic costs by only attending over a small weighted coreset. Crucially, we select the coreset using a fast but spectrally-accurate subsampling algorithm -- randomly pivoted Cholesky -- and weight the elements optimally to minimise reconstruction error. Remarkably, given bounded inputs, WildCat approximates exact attention with super-polynomial $O(n^{-\sqrt{\log(\log(n))}})$ error decay while running in near-linear $O(n^{1+o(1)})$ time. In contrast, prior practical approximations either lack error guarantees or require quadratic runtime to guarantee such high fidelity. We couple this advance with a GPU-optimized PyTorch implementation and a suite of benchmark experiments demonstrating the benefits of WildCat for image generation, image classification, and language model KV cache compression.
Science Journals
arXiv:2602.10120v2 Announce Type: replace
Abstract: We provide a numerical method for computing solutions to a free boundary problem arising from the equilibrium state of a floating drop. This numerical method is based on a Newton's method for the underlying nonlinear boundary value problems, and at each iterative step a Chebyshev spectral collocation method is employed. The problems considered here are those that can be described by using generating curves, and include problems in $\mathbb{R}^2$ and $\mathbb{R}^3$.
The resulting nine-dimensional space of physical parameters is explored, and examples are given that highlight the potential energy of centrally located drops, wall-bound drops, and asymmetrical configurations in $\mathbb{R}^2$. Non-uniqueness of solutions to the corresponding Euler-Lagrange equations is displayed, and also strong evidence of non-uniqueness of energy minimizers is given.
arXiv:2602.10623v2 Announce Type: replace
Abstract: Reward models learned from human preferences are central to aligning large language models (LLMs) via reinforcement learning from human feedback, yet they are often vulnerable to reward hacking due to noisy annotations and systematic biases such as response length or style. We propose Bayesian Non-Negative Reward Model (BNRM), a principled reward modeling framework that integrates non-negative factor analysis into Bradley-Terry (BT) preference model. BNRM represents rewards through a sparse, non-negative latent factor generative process that operates at two complementary levels: instance-specific latent variables induce disentangled reward representations, while sparsity over global latent factors acts as an implicit debiasing mechanism that suppresses spurious correlations. Together, this disentanglement-then-debiasing structure enables robust uncertainty-aware reward learning. To scale BNRM to modern LLMs, we develop an amortized variational inference network conditioned on deep model representations, allowing efficient end-to-end training. Extensive empirical results demonstrate that BNRM substantially mitigates reward over-optimization, improves robustness under distribution shifts, and yields more interpretable reward decompositions than strong baselines.
arXiv:2606.01234v1 Announce Type: cross
Abstract: The GDP of a country is modelled as the relative interaction between two agents - working hours, reflecting the social choice of a population, and Total Factor Productivity, reflecting the collective investment in productivity enhancers. It is shown that a Random Forest model can accu- rately predict the GDP from these two factors. The differences in the choices made by Germany and USA are analysed though Gini importance, SHAP plots and partial dependency. It is shown that the differences in the social structure of the countries are reflected in the relative contribution of working hours and productivity to the GDP.
arXiv:2509.20070v2 Announce Type: replace
Abstract: We present LLM Trainer, a fully automated pipeline that leverages the world knowledge of Large Language Models (LLMs) to transform a small number of human demonstrations (as few as one) into a large robot dataset for imitation learning. Our approach decomposes demonstration generation into two steps: (1) offline demonstration annotation that extracts keyframes, salient objects, and pose-object relations; and (2) online keypose retargeting that adapts those keyframes to a new scene, given an initial observation. Using these modified keypoints, our system warps the original demonstration to generate a new trajectory, which is then executed, and the resulting demo, if successful, is saved. Because the annotation is reusable across scenes, we use Thompson sampling to optimize the annotation, significantly improving generation success rate. We evaluate our method on a range of tasks, and find that our data annotation method consistently outperforms expert-engineered baselines. We further show an ensemble policy that combines the optimized LLM feed-forward plan with a learned feedback imitation learning controller. Finally, we demonstrate hardware feasibility on a Franka Emika Panda robot. For additional materials and demonstration videos, please see the project website: https://sites.google.com/andrew.cmu.edu/llm-trainer
arXiv:2509.21474v4 Announce Type: replace
Abstract: While diffusion language models (DLMs) have achieved competitive performance in text generation, improving their reasoning ability with reinforcement learning remains an active research area. Here, we introduce d2, a reasoning framework tailored for masked DLMs. Central to our framework is a new policy gradient algorithm that relies on accurate estimates of the sampling trajectory likelihoods. Because computing these likelihoods naively is computationally expensive for masked DLMs, we develop a family of estimators tailored to distinct model classes. For DLMs that support a sampling algorithm called any-order decoding, we propose d2-AnyOrder, which achieves exact trajectory likelihood with a single model pass. Through an empirical study of widely used DLMs, we show that any-order decoding is not universally supported in practice. For standard masked diffusion models, we propose d2-StepMerge, which approximates the trajectory likelihood, trading off compute for approximation accuracy in an analytically tractable manner. Empirically, d2 significantly outperforms widely-used RL baselines when applied to popular DLMs, and sets a new state-of-the-art performance for DLMs on logical reasoning tasks (Countdown and Sudoku) and math reasoning benchmarks (GSM8K and MATH500). We provide the code along with a blog post on the project page: https://guanghanwang.com/d2
arXiv:2503.11832v5 Announce Type: replace
Abstract: Recent vision language models (VLMs) have made remarkable strides in generative modeling with multimodal inputs, particularly text and images. However, their susceptibility to generating harmful content when exposed to unsafe queries raises critical safety concerns. While current alignment strategies primarily rely on supervised safety fine-tuning with curated datasets, we identify a fundamental limitation we call the ''safety mirage'', where supervised fine-tuning inadvertently reinforces spurious correlations between superficial textual patterns and safety responses, rather than fostering deep, intrinsic mitigation of harm. We show that these spurious correlations leave fine-tuned VLMs vulnerable even to a simple one-word modification-based attack, where substituting a single word in text queries with a spurious correlation-inducing alternative can effectively bypass safeguards. Additionally, these correlations contribute to the over-prudence, causing fine-tuned VLMs to refuse benign queries unnecessarily. To address these issues, we show machine unlearning (MU) as a powerful alternative to supervised safety fine-tuning, as it avoids biased feature-label mappings and directly removes harmful knowledge from VLMs while preserving their general capabilities. Extensive evaluations across safety benchmarks show that under MU-based alignment reduces the attack success rate by up to 60.27% and cuts unnecessary rejections by over 84.20%. WARNING: There exist AI generations that may be offensive in nature.
arXiv:2506.09220v2 Announce Type: replace
Abstract: A new generation of AI personal assistants reached consumers in 2023-2024 amid sweeping claims about anticipatory, agentic intelligence. Wearables such as the Rabbit R1 and Humane AI Pin, and subscription services such as Ohai and Docus, promised to learn users' routines and complete tasks across digital platforms. Drawing on semi-structured interviews with nine early adopters, this article asks how users make sense of these systems when the imaginary of an autonomous "second self" meets the recalcitrance of actual devices. Extending uncertainty reduction theory, we specify three forms of uncertainty in initial encounters: functional (what can it do?), relational (how do I get it to do it?), and metaphysical (what is it to me, and what should it remember?). We find that hype continues the pre-domestication of voice assistants; that the most satisfying uses are user-curated constellations of narrow tools rather than standalone "second selves."
arXiv:2606.00393v1 Announce Type: cross
Abstract: Geometric distortion in prostate diffusion-weighted imaging (DWI) can impair lesion localization and reduce the reliability of MRI-based clinical assessment. We propose AutoIQ, an ensemble machine learning framework for automatic quantification and classification of DWI geometric distortion severity. A total of 140 retrospective prostate biparametric MRI examinations were analyzed, including 33 scans with severe distortion requiring repeat acquisition and 107 scans with acceptable distortion based on expert radiologist assessment. AutoIQ combines two complementary distortion quantification strategies: a segmentation-based method measuring prostate boundary mismatch between T2-weighted imaging (T2WI) and DWI, and a registration-based method estimating deformation magnitude after DWI-to-T2WI alignment. The resulting distortion scores were used to train individual classifiers and a logistic-regression ensemble model. Both computational methods significantly differentiated severe from acceptable distortion cases (p < 0.001). On an independent test set, the ensemble model achieved an accuracy of 0.95, F1-score of 0.93, and AUC of 0.98, outperforming individual models. These results suggest that AutoIQ can provide automated, quantitative quality assessment for prostate DWI and may help identify scans that require repeat acquisition.
arXiv:2606.00403v1 Announce Type: cross
Abstract: Predicting how crystal structure influences high-temperature stability remains a key challenge in materials modelling and design. Silicon carbide (SiC), one of the most thermally and chemically stable materials known, provides an ideal system for studying this problem because its many polytypes preserve similar local tetrahedral bonding while differing in long-range stacking geometry. Here, we combine phase-coexistence machine-learning molecular dynamics with finite-temperature phonon analysis, enabled by a fine-tuned MACE interatomic potential that accurately describes crystalline, high-temperature, and disordered configurations across multiple SiC polytypes. We identify a clear relative stability ordering, 3C > 2H > 9R, reflected consistently in structural disordering, interlayer sliding, and finite-temperature phonon spectra. Across all polytypes, melting initiates through the formation of short C-C contacts and carbon-rich local regions, followed by a progressive loss of tetrahedral Si-C connectivity. The reduced stability of the long-period 9R polytype is traced to low-frequency transverse-acoustic shear modes associated with relative bilayer sliding, which are already present in the 0 K phonon spectra and soften further at high temperature. These modes generate larger lateral bilayer displacements, linking enhanced interlayer sliding to local chemical disordering and ultimately melting. More broadly, our results show that high-temperature stability in polytypic covalent materials is governed not only by local bond strength, but also by stacking-dependent transverse dynamics.
arXiv:2606.00409v1 Announce Type: cross
Abstract: Unconventional antiferromagnets combine zero net magnetization with spin-split electronic bands, offering a distinct, important platform for spintronics. Their discovery, however, has so far depended largely on case-by-case studies and on a limited number of compounds with experimentally resolved magnetic structures. Here, we overcome these bottlenecks by resolving magnetic ground states across a broad materials database. We narrow down 37163 magnets from the Materials Project to 189 collinear antiferromagnets by combining physics-informed prescreening, high-throughput exchange calculations and Luttinger-Tisza analysis. Among these, symmetry analysis identifies 36 altermagnets and 11 Luttinger-compensated ferrimagnets (LCFs), including 22 altermagnets and 9 LCFs that have not been reported previously. The identified unconventional antiferromagnets can support nonrelativistic spin Hall effects and doping-tunable spin transport with switchable polarization and giant anisotropy. Our framework converts broad structural databases into a curated, symmetry-classified set of experimentally testable compensated spin-split magnets, establishing a scalable route for the efficient discovery of functional antiferromagnets.
arXiv:2606.00413v1 Announce Type: cross
Abstract: Sufficient dimension reduction (SDR) makes high-dimensional regression tractable by projecting the covariates onto a low-dimensional subspace that preserves the conditional mean of the response. Existing gradient-based estimators either operate in the ambient space and suffer from the curse of dimensionality, or localize in the reduced space at a per-outer-iteration cost at least quadratic in the sample size. We show that minimizers of the population Minimum Average Variance Estimation (MAVE) risk approximate the same Grassmannian target as the Outer Product of Gradients (OPG), and recast the empirical criterion as a smooth maximization on the Stiefel manifold with closed-form Riemannian gradient. The resulting algorithm, SMAVE, combines sparse projected-space nearest-neighbor localization with Riemannian stochastic gradient ascent. A simplified version comes with almost-sure convergence and a non-asymptotic rate matching the standard non-convex stochastic first-order scaling. Empirically, SMAVE matches or improves on RMAVE's synthetic subspace recovery at moderate-to-high ambient dimension, and on four real datasets it uniformly improves over OPG and is competitive with or outperforms RMAVE at orders of magnitude lower runtime.
arXiv:2602.11554v3 Announce Type: replace
Abstract: How far can 3D object detection go using 4D radar alone? Despite offering weather-robust and velocity-aware sensing for autonomous perception, modern 4D radar still yields sparse, noisy, and unstable point clouds, limiting radar-only 3D detection. We present HyperDet, a detector-agnostic framework that constructs task-aware hyper 4D radar point clouds before detection. HyperDet first refines short-window surround-view radar observations through spatio-temporal accumulation, cross-sensor validation, and Doppler-guided motion compensation, improving return reliability and temporal coherence. It then performs foreground generative enhancement using LiDAR-guided pseudo-radar supervision available only during training, enriching object geometry while preserving measured radar background and radar-native attributes. During detector training, radar-aware object-level augmentation further preserves Doppler consistency under geometric relocation. At inference time, HyperDet requires radar input alone and can be directly paired with standard 3D detectors. Experiments on two public surround-view 4D radar datasets demonstrate consistent improvements over raw radar inputs across standard 3D detectors, validating input-level radar enhancement as an effective approach to radar-only 3D detection.
arXiv:2602.13937v2 Announce Type: replace
Abstract: Automated Machine Learning (AutoML) has improved access to machine learning, yet existing techniques often remain limited in flexibility, transparency, and execution reliability. Code-driven AutoML offers a promising direction by synthesizing executable code for preprocessing, model training, and evaluation. However, current LLM-based approaches frequently generate code that is plausible in text yet brittle in execution, insufficiently grounded in the actual dataset, or restricted to narrow solution paths. In this paper, we introduce iML, a multi-agent code-driven AutoML framework designed around three requirements: executability, problem grounding, and broad exploration of valid solutions. iML first analyzes the task and profiles the data, then synthesizes a structured blueprint that guides modular code generation across multiple implementation tracks, including traditional ML,pretrained adaptation, and custom neural architectures. To improve reliability, iML enforces interface checking, dynamic execution, and iterative debugging during integration. We evaluate iML on MLE-BENCH and the newly introduced iML-BENCH, covering diverse Kaggle-style tasks. On MLE-BENCH, iML attains a 90% valid submission rate and a 45% medal rate, and an APS of 0.82, improving the average standardized performance score (APS) over the LLM-based baselines by 52%-273%. On iML-BENCH, it achieves the highest APS and demonstrates robust performance even when task descriptions are substantially stripped. These results establish iML as a reliable and competitive framework for code-driven AutoML.
arXiv:2602.14065v2 Announce Type: replace
Abstract: Knowledge-intensive Visual Question Answering (KI-VQA) frequently suffers from severe knowledge conflicts caused by the inherent limitations of open-domain retrieval. However, existing paradigms face critical limitations due to the lack of generalizable conflict detection and intra-model constraint mechanisms to handle conflicting evidence. To address these challenges, we propose the REAL (Reasoning-Pivot Alignment) framework centered on the novel concept of the Reasoning-Pivot. Distinct from reasoning steps that prioritize internal self-derivation, a reasoning-pivot serves as an atomic unit (node or edge) in the reasoning chain that emphasizes knowledge linkage, and it typically relies on external evidence to complete the reasoning. Supported by our constructed REAL-VQA dataset, our approach integrates Reasoning-Pivot Aware SFT (RPA-SFT) to train a generalizable discriminator by aligning conflicts with pivot extraction, and employs Reasoning-Pivot Guided Decoding (RPGD), an intra-model decoding strategy that leverages these pivots for targeted conflict mitigation. Extensive experiments on diverse datasets demonstrate that REAL significantly enhances discrimination accuracy and achieves superior performance, validating our pivot-driven resolution paradigm.
arXiv:2602.14134v2 Announce Type: replace
Abstract: Multimodal Large Language Models (MLLMs) have demonstrated exceptional capabilities in high-level visual understanding. However, extending these models to fine-grained dense prediction tasks, such as semantic segmentation and depth estimation, typically necessitates the incorporation of complex, task-specific decoders and other customizations. This architectural fragmentation increases model complexity and deviates from the generalist design of MLLMs, ultimately limiting their practicality. In this work, we challenge this paradigm by accommodating standard MLLMs to perform dense predictions without requiring additional task-specific decoders. The proposed model is called DenseMLLM, grounded in the standard architecture with a novel vision token supervision strategy for multiple labels and tasks. Despite its minimalist design, our model achieves highly competitive performance across a wide range of dense prediction and vision-language benchmarks, demonstrating that a standard, general-purpose MLLM can effectively support dense perception without architectural specialization. This project is available at github.com/Eli-YiLi/DenseMLLM.
arXiv:2606.00483v1 Announce Type: cross
Abstract: Genotype-based cis-expression prediction depends on accurately modeling local regulatory architecture. We present block-sparse Bayesian sparse linear mixed model (bsBSLMM), an extension of Bayesian sparse linear mixed model (BSLMM) that incorporates linkage disequilibrium (LD)-block spike-and-slab sparsity and a transcription start site (TSS)-informed SNP inclusion prior. Across 23,098 genes from GEUVADIS European-ancestry lymphoblastoid cell lines, bsBSLMM retained more predictable genes than BSLMM, LASSO, BLUP, TIGAR elastic net, and TIGAR Dirichlet-process regression under matched evaluation criteria. Compared with BSLMM, bsBSLMM improved held-out prediction performance for most shared genes, with gains driven primarily by LD-block sparsity and further enhanced by the TSS-informed prior. Variants selected by bsBSLMM showed stronger enrichment in GM12878 DNase and H3K27ac regulatory regions than variants selected by BSLMM. In transcriptome-wide association study (TWAS) analysis, bsBSLMM recovered established inflammatory bowel disease signals, including IL23R, and identified additional genome-wide significant genes not detected by BSLMM. Independent validation in the Louisiana Osteoporosis Study reproduced the increased prediction yield across ancestries and recovered biologically relevant bone mineral density pathways in downstream TWAS and gene set enrichment analyses. These results demonstrate that incorporating LD-block structure and biologically informed SNP priors improves cis-expression prediction and enhances downstream TWAS discovery.
arXiv:2606.00495v1 Announce Type: cross
Abstract: Two-dimensional conjugated polymers (2DCPs) are significant macromolecular materials with intriguing and tunable physicochemical properties that depend on their geometries. Graphdiyne and its derivatives are exemplary 2DCPs featuring sp-sp2 hybridized skeletons. However, achieving single-layered, large-domain/regular graphdiyne and its derivatives on surfaces remains a formidable challenge due to the lack of selective 2D covalent polymerization methods. Here, we report a selective on-surface 2D covalent polymerization method via the combination of cobalt catalysis and coronene templating, achieving the synthesis of single-layered fluorographdiyne nanosheets up to 60*60 nm2 on Au(111) surface. Using scanning probe techniques, we visualize the sequential polymerization process and characterize cobalt-activated coupling intermediates at the atomic level. Experimental and theoretical analyses suggest that strong d-{\pi} coupling between cobalt and alkynyl transforms a robust Csp-Au bond into a weaker Csp2-Au bond, thereby facilitating the demetallization C-C coupling. Besides, the templating effect of coronene suppresses kinetically trapped defects and improves the selectivity of hexagonal-ring formation in the complex 2D covalent polymerization process.
arXiv:2606.00501v1 Announce Type: cross
Abstract: We consider an agent, who would like to execute a given quantum circuit using resources leased from a set of quantum computers (QCs) connected by a quantum network. For this purpose, the agent needs to make the following four key decisions: (i) how many qubits to lease from each QC, (ii) at which QCs to store different circuit qubits in different time slots, (iii) at which QC to execute each gate in the circuit, and (iv) how to move qubits between QCs, choosing between migration and teleportation. We refer to this problem facing the agent as the joint qubit leasing and quantum circuit distribution (JQLQCD) problem, and provide a comprehensive integer linear programming (ILP) formulation for it. We show that the JQLQCD problem is NP-complete. Next, we identify several special cases in which the problem can be optimally solved in closed form or via polynomial-time algorithms. Also, we propose a greedy algorithm with local search refinement to solve large instances of the general JQLQCD problem. Finally, we evaluate the performance of the proposed greedy algorithm using extensive numerical computations.
arXiv:2606.00521v1 Announce Type: cross
Abstract: The relativistic transformation of temperature has long remained controversial, with the classical laws of Planck-Einstein, Ott-Eddington-Moller and Landsberg yielding conflicting results. We reexamine this issue from a relativistic thermodynamic and statistical perspective, starting from the energy-momentum tensor of an isotropic system and defining the effective temperature Teff as that inferred by a moving observer from the transformed energy density. Analyses of a photon gas, a relativistic ideal gas and an electron gas show that Teff consistently increases with velocity, supporting the Ott-Eddington interpretation while depending on the system's equation of state. These results indicate that temperature is not a Lorentz-invariant scalar but an observer-dependent quantity. A consistent relativistic description emerges when temperature is related to the inverse-temperature four-vector beta, linking operational and invariant viewpoints within a unified thermodynamic framework.
arXiv:2606.00584v1 Announce Type: cross
Abstract: This paper proposes Spectra-Guided Neural Tucker Factorization (SG-NTF) for High-Dimensional and Incomplete (HDI) tensor completion. Circumventing discrete representational limits, SG-NTF maps scalar timestamps into a continuous spectral space to abstract temporal periodicities. Concurrently, a Spatio-Temporal Co-Gating (STCG) mechanism explicitly filters latent interactions via multiplicative modulation on spatiotemporal contexts. Evaluations on real-world HDI tensors verify that SG-NTF maintains competitive completion accuracy with parameter efficiency.
arXiv:2606.00615v1 Announce Type: cross
Abstract: Weyl semimetals obey a global chirality constraint: the net chiral topological charge and any associated chiral spectral flow must vanish, as required by the Nielsen-Ninomiya theorem. Under magnetic fields, this constraint manifests through counter-propagating zeroth Landau levels associated with Weyl nodes of opposite chirality. Here, we experimentally demonstrate how non-Hermiticity can reshape this balance in a synthetic photonic Weyl semimetal. Using engineered gauge fields in one-dimensional multilayer structures, we realize both homogeneous and axial magnetic fields and directly probe the resulting Landau-level spectra. While a homogeneous field produces the expected chirality-balanced zeroth Landau levels, an axial field spatially separates the compensating chiral channels: co-propagating bulk pseudo-Landau levels carry one chirality, whereas the opposite chirality resides in boundary-localized surface states. We show that radiative boundary loss selectively suppresses these surface states, removing them from the long-lived observable spectrum and producing an experimentally accessible chirality imbalance. By reducing boundary loss, we recover the hidden chiral channel and reveal its surface-state origin. These results show that non-Hermiticity, present naturally in photonics, can control and relax fundamental chirality constraints in topological systems, enabling access to otherwise forbidden spectral responses.
arXiv:2606.00665v1 Announce Type: cross
Abstract: Sliding ferroelectricity in van der Waals (vdW) layered systems has emerged as a promising route toward non-volatile nanoscale devices, where interlayer displacement in non-centrosymmetric bilayers generates an out-of-plane polarization. In particular, 3R-stacked bilayer transition metal dichalcogenides (TMDs) grown via chemical vapor deposition (CVD) have been shown to host such polarization due to broken inversion symmetry. However, a detailed investigation of the 2D ferroelectric (FE) properties of CVD-grown 2D films, particularly the role of intrinsic disorder, such as structural defects and domain structure, remains poorly understood. Here, we investigate the FE switching characteristics of CVD-grown 3R-stacked WSe2 using a graphene-based ferroelectric field-effect transistor (graphene-FE-FET) architecture, where graphene serves as a highly sensitive probe of induced charge modulation due to polarization switching of FEs. We show that the growth-induced structural disorder significantly impacts polarization switching, while multi-domain kinetics governs the evolution of the FE response. These findings provide important insights into the design and optimization of FE devices based on vdW materials.
arXiv:2505.11791v2 Announce Type: replace
Abstract: To steer the behavior of selfish, resource-sharing agents in a socio-technical system towards the direction of higher efficiency, the system designer requires accurate models of both agent behaviors and the underlying system infrastructure. For instance, traffic controllers often use road latency models to design tolls whose deployment can effectively mitigate traffic congestion. However, misspecifications of system parameters may restrict a system designer's ability to influence collective agent behavior toward efficient outcomes. In this work, we study the impact of system misspecifications on toll design for atomic congestion games. We prove that tolls designed under sufficiently minor system misspecifications, when deployed, do not introduce new Nash equilibria in atomic congestion games compared to tolls designed in the noise-free setting, implying a form of local robustness. We then upper bound the degree to which the worst-case equilibrium system performance could decrease when tolls designed under a given level of system misspecification are deployed. We validate our theoretical results via Monte-Carlo simulations as well as realizations of our worst-case guarantees.
arXiv:2606.00666v1 Announce Type: cross
Abstract: Transition metal complexes are central to catalysis, drug design, and materials science, with relevant properties strongly sensitive to their three-dimensional geometry. However, the electronic diversity and unconventional bonding environments of transition metal complexes pose a major challenge for accurate structure generation. In this work, we introduce TMCgen, a manifold diffusion machine learning model that efficiently and accurately generates geometries of transition metal complexes. By formulating the diffusion process over the metal-ligand coordination angles, combined with torsional and rotational diffusion of the ligands, TMCgen focuses on the key geometric degrees of freedom of transition metal complexes. TMCgen shows strong performance in generating accurate coordination environments on a diverse set of experimentally derived bioinorganic and organometallic complexes while requiring only few inference steps, enabling efficient generation. Our results demonstrate the potential of manifold-based generative modeling for data-efficient geometry generation, paving the way for property-conditioned design of transition metal complexes.