arXiv subjects

Yusong Zhao

Publications and source records attributed to Yusong Zhao.

8 recordsLinked to original sources

arXiv2026

ResemBrick: Brick Reconstruction from Photographs with Perceptual Fidelity and Buildability

Producing a hand-buildable, colored brick model of a 3D object from a few casual photographs is a clean testbed for a broader challenge: generating 3D content that meets hard physical-assembly constraints under a discrete, budget-limited voxel grid. On a coarse lattice, visual resemblance and structural stability pull against each other, yet prior brick pipelines address only one side and treat voxelization as fixed preprocessing rather than a variable to optimize. We present ResemBrick, which couples the two. Budgeted occupancy completion reframes discretization as allocation: given a target occupied-voxel count, a single resolution-conditioned network decides in one feed-forward pass which surface voxels to fill for best appearance, one weight set spanning 13 resolutions. Buildability by construction then combines support- and look-ahead-aware greedy placement with a deterministic, provably terminating repair that grounds every floating component. Under a matched budget, ResemBrick surpasses existing voxel selectors in perceptual fidelity while uniquely reaching zero floating and zero unstable bricks on unfiltered held-out objects; as a complete pipeline, it attains the best perceptual fidelity among prior brick-construction systems. Our results point to treating discretization and assembly as tightly coupled stages rather than independent ones.

cs.CV↗

arXiv2026

Interpretable GOHR Agents via Sparse Autoencoders

A central challenge in interpreting learned decision-making systems is to determine whether their internal representations contain concepts that help explain their behavior. We report interpretability experiments for a tokenized autoregressive Transformer agent in the Game of Hidden Rules (GOHR). We focus on a compact two-rule task in which both hidden rules map object shapes to target buckets, but with different permutations. The policy is trained on episodes sampled from these two hidden rules and then evaluated with fixed weights. It is never given a rule label and does not use an explicit rule classifier; any rule information must be inferred implicitly from interaction history. In this setting, the correct rule is not identifiable before the agent tries an informative move and observes accept/reject feedback. Sparse autoencoders (SAEs) trained on the agent's decision-token embeddings recover this structure. When held-out decisions are labeled by simple concepts such as the chosen shape or bucket, SAE dimensions that are highly selective for a concept cover most decisions where that concept is present. Individual SAE dimensions also correspond to interpretable strategies such as probing one rule hypothesis and switching after negative feedback.

cs.LG↗

arXiv2026

Probabilistic Residual Learning for Online Recommendations

Modern recommender systems are typically based on deep learning (DL) models, where a dense encoder learns representations of users and items. As a result, these systems often suffer from the black-box nature and computational complexity of the underlying models, making it difficult to systematically enhance their recommendation capabilities. To address this problem, we propose Probabilistic Residual Learning (PRL), a causal Bayesian recommendation model that models the residual between ground-truth and base predictions, enabling targeted refinement of existing systems. Specifically, PRL (1) probabilistically groups users for localized residual modeling, (2) models domain-level confounders that influence user and item representations, and (3) aggregates cluster-specific residual predictions over the confounders using do-calculus. Experiments demonstrate that our plug-and-play PRL is compatible with various base deep learning recommender systems, improving their performance while automatically discovering meaningful user clusters.

cs.IR↗

arXiv2026

Counterintuitive inverse superconducting transition beyond 4He-cooling limit

Thermally driven quantum-orders observed in exceptional instances may redefine the role of thermal-fluctuation from a source of decoherence to a resource for coherent-state engineering. While preliminary signs of counterintuitive temperature-rise-triggered superconductivity manifested in CeCu2Si2, ErRh4B4, Ho1.2Mo6S8 and (La,Ce)Al2, their critical-temperatures (Tc-inv) remain below Kelvin-range, precluding substantial applications. Here, we report field-modulated inverse-superconducting-transitions above 4He-cooling-limit in Eu-based infinite-layer nickelates (EuxNd1-xNiO2 and EuxPr1-xNiO2) grown on a substrate under both overdoped and underdoped regimes. Paradigmatically, superconductivity with zero-resistance is confined between Tc-inv (2.6-5.4 K) and another higher normal-Tc, rising and decreasing with applied magnetic-field, respectively. Starting from the resistive-state below Tc-inv, the inverse-superconducting-transition is driven by not only temperature-rising, but also current-density, while superconductivity further vanishes at higher temperature and current thresholds. The Kelvin-range inverse superconducting transition is plausibly explained by temperature-induced alternating dominance of effective magnetic-fields arising from Eu2+4f7 related compensations relative to the upper-critical-field. Furthermore, an extended-phenomenological-framework is also supported by reemerged superconductivity below 300 mK under magnetic-field, giving rise to an unprecedented temperature-induced reentrant superconductivity. Our findings establish magnetic-interaction-reconfigured high-Tc systems as fertile platforms for exploring quantum phenomena that reverse thermal-decoherence paradigm, also enabling antithetical-designs to unlock untapped application-scenarios for quantum-phase-transition devices.

cond-mat.supr-con↗

arXiv2026

Cascaded Sparse Autoencoders Learn Multi-Level Visual Concepts in Multimodal LLMs

Multimodal Large Language Models (MLLMs) have demonstrated strong performance on vision-language tasks, yet their internal visual representations remain difficult to interpret. Sparse Autoencoders (SAEs) provide a scalable way to decompose dense model activations into sparse, interpretable features. However, existing SAE architectures primarily recover flat feature dictionaries and are less suited for explicit multi-level concept organization. In this paper, we introduce cascaded sparse autoencoders (CSAEs) for learning hierarchical visual concepts in MLLMs. Rather than nesting or stacking SAE sparse activation codes, CSAEs train a second-level SAE directly on the decoder weights of the first-level SAE, treating learned low-level feature directions as inputs for higher-level abstraction. This design enables CSAEs to learn "concepts of concepts" while avoiding drawbacks from the shared-prefix coupling of nesting, Matryoshka-style hierarchies and the bottlenecks of naively stacked SAEs. Experiments across Qwen3-VL, Gemma-3, and LLaVA on multiple visual datasets show that CSAEs improve interpretability in terms of hierarchical concept coherence over state-of-the-art SAE baselines. Results on concept steering further demonstrate that the learned concept groups support effective group-level interventions in MLLM outputs.

cs.CV↗

arXiv2026

PaSBench-Video: A Streaming Video Benchmark for Proactive Safety Warning

Between the first visible sign of danger and the moment an accident occurs, there is often a window where intervention remains possible. Video-capable multimodal large language models (MLLMs) could serve as always-on safety monitors that issue warnings during this window. Yet current benchmarks do not test this ability: they rely on static inputs, ignore timing precision, and omit false-positive measurement on safe scenes. We present PaSBench-Video, a 740-video benchmark with 481 risk and 259 no-risk videos across four domains: driving, healthcare, daily life, and industrial production. Risk videos are annotated with frame-level risk onset and accident boundaries. A model must observe the video causally and produce a warning that is both temporally calibrated and content-correct. Testing 13 MLLMs, we find that no model exceeds 20.0% on our strictest metric, and recall is tightly coupled with false-positive rate, with Pearson correlation 0.64: higher detection comes only at the cost of triggering warnings on the majority of safe clips. Performance splits sharply by domain: models achieve moderate recall at low false-positive rates in daily life, where risks are inherently anomalous, yet fire indiscriminately in driving, where routine and hazardous scenes look alike. These results indicate that current models rely on scene-level activity cues rather than reasoning about emerging harm.

cs.CL↗

arXiv2025

A chemical avenue to manipulate field-reentrant superconducting rivalries in infinite layer nickelates

Recently, preliminary magnetic field-reentrant superconductivity manifested in high-temperature (Tc) Eu-doped infinite-layer (IL) nickelates, beyond analogous discoveries exclusively in low-Tc systems. This evokes intriguing fundamental issues about potential quantum-phase boundary and criticality between unconventional superconductivity and field-reentrant-one, which are inexplicable owing to formidable challenges in growing IL-nickelates towards later-series rare-earths. Herein, we open up chemical avenues to enable effective growth of (RE1-yRE'y)1-xEuxNiO2 (RE/RE': Pr, Nd, Sm, Gd, Dy), giving rise to discoveries of RE-4f-related quantum competition between high-Tc and reentrant superconductivity. Robust magnetic-field-reentrant superconductivity with uniaxial anisotropy is observed at superconducting-dome boundaries, stemming from Eu2+-4f7 associated competition between magnetic-fluctuation promoted pairing and exchange-field interactions. Their quantum-criticality is further modulable via RE(RE')-magnetism, which either reinforces reentrancy or elevates Tc (40.1 K) with more robust critical-current-density (~266 kA/cm2 at 2 K) beyond Sr-/Ca-doped counterparts. Our synthetic route enables the establishment of an ideal platform via IL-nickelates for studying 4f-related unconventional superconductivity and quantum-criticality.

cond-mat.supr-con↗

arXiv2021

Scalable Multi-Task Gaussian Processes with Neural Embedding of Coregionalization

Multi-task regression attempts to exploit the task similarity in order to achieve knowledge transfer across related tasks for performance improvement. The application of Gaussian process (GP) in this scenario yields the non-parametric yet informative Bayesian multi-task regression paradigm. Multi-task GP (MTGP) provides not only the prediction mean but also the associated prediction variance to quantify uncertainty, thus gaining popularity in various scenarios. The linear model of coregionalization (LMC) is a well-known MTGP paradigm which exploits the dependency of tasks through linear combination of several independent and diverse GPs. The LMC however suffers from high model complexity and limited model capability when handling complicated multi-task cases. To this end, we develop the neural embedding of coregionalization that transforms the latent GPs into a high-dimensional latent space to induce rich yet diverse behaviors. Furthermore, we use advanced variational inference as well as sparse approximation to devise a tight and compact evidence lower bound (ELBO) for higher quality of scalable model inference. Extensive numerical experiments have been conducted to verify the higher prediction quality and better generalization of our model, named NSVLMC, on various real-world multi-task datasets and the cross-fluid modeling of unsteady fluidized bed.

stat.ML↗