arXiv subjects

Shixiong Xu

Publications and source records attributed to Shixiong Xu.

9 recordsLinked to original sources

arXiv2026

Anomalies, higher symmetries, and homotopy theory I: crossed precosheaves and a nonabelian Cech construction

Symmetries of a quantum field theory or a quantum lattice model are expected to be described not by a group but by a homotopy type, whose homotopy groups are the higher symmetry groups and whose Postnikov invariants encode 't Hooft anomalies. We propose a homotopy-theoretic definition of 't Hooft anomalies based on the structure of localized symmetries in the operator-algebraic approach. The input is a crossed precosheaf: a collection of groups attached to regions of space, equipped with pairings recording that the commutator of two transformations localized in two regions is localized in their intersection. To a crossed precosheaf and a finite family of regions we attach a pointed connected homotopy type, the Cech-Loday classifying space. As the family of regions ranges over finite covers of space, the classifying spaces form a pro-object in the homotopy category. There is a canonical map from the ordinary classifying space of the symmetry group to this pro-type. Composing this map with a group action gives an anomaly whose nontriviality obstructs gauging. In favorable cases, the Cech-Loday space is essentially independent of the cover. In such cases, the homotopy fiber of the canonical map is a candidate for the classifying space of the higher group of symmetries. We give examples of crossed precosheaves arising from gauge transformations in continuum field theory and from quantum cellular automata and finite-depth circuits. For quantum spin systems in one and two dimensions, our construction recovers the Else-Nayak index and its generalizations to two dimensions.

math-ph↗

arXiv2026

PACE: Temporal-Role Tactile Conditioning for Chunk-Based Insertion Policies

Contact-rich insertion requires robots to combine visual alignment with tactile adjustment. However, existing visuotactile chunk-based policies collapse tactile feedback into a single generation-time fusion pathway, leaving persistent contact states, abrupt contact events, and future contact consequences implicitly modeled in a single representation. We present PACE (Phase-Aware Contact-Event Conditioning), a temporal-role tactile conditioning framework for chunk-based insertion policies. PACE organizes a shared tactile stream into three temporal roles: persistent contact states condition action-chunk generation, abrupt contact events correct the generated chunk, and contact-consequence prediction shapes action representations during training. This design preserves chunk-level action continuity while assigning distinct pathways to different temporal roles of contact feedback. Across five real-robot insertion tasks, PACE achieves an average success rate of 90.7%, outperforming vision-only and generation-time tactile-fusion baselines by over 44 percentage points; on the ManipulationNet benchmark, PACE maintains 88.7% success across six geometry-clearance configurations, including an industrial-grade 0.1-mm clearance. Ablations confirm the complementary contributions of the three temporal roles, mechanistic analyses show that PACE separates normal insertion progress from contact events requiring immediate correction, and cross-signal and cross-architecture experiments indicate that this temporal-role principle extends across different physical feedback signals and policy backbones. Videos and materials are available at https://pace-insertion.github.io/.

cs.RO↗

arXiv2026

Finder: Agentic Closed-Loop Object Finding for Embodied Grounding

Finding the object referred to by language in a partially observed 3D scene is a core capability for embodied agents. Existing approaches either couple object search with online exploration, which can be costly when relevant observations have already been captured, or query pre-built open-vocabulary maps and scene graphs in a static, one-shot fashion. We present Finder, an agentic closed-loop object-finding primitive for embodied grounding. Instead of treating grounding as passive retrieval from a fixed scene representation, Finder maintains a typed loop state that links query-conditioned planning, scoped evidence gathering, candidate verification, and accept/continue/abort control. When evidence is incomplete or ambiguous, the loop can redirect subsequent perception and comparison rather than simply returning the top retrieved object. On open-vocabulary embodied Object Retrieval in Habitat/HM3D and real-world RGB-D scenes, Finder improves the averaged 1m success rate by 15.75 points over strong baselines. The same primitive also transfers to sequential object grounding and embodied object-centric question answering, improving spatial and temporal localization without changing the inner grounding protocol. Project page: https://finder-vln.github.io.

cs.CV↗

arXiv2025

Higher symmetries and anomalies in quantum lattice systems

We define an 't Hooft anomaly index for a group acting on a 2d quantum lattice system by finite-depth circuits. It takes values in the degree-4 cohomology of the group and is an obstruction to on-siteability of the group action. We introduce a 3-group (modeled as a crossed square) describing higher symmetries of a 2d lattice system and show that the 2d anomaly index is an obstruction for promoting a symmetry action to a morphism of 3-groups. This demonstrates that 't Hooft anomalies are a consequence of a mixing between ordinary symmetries and higher symmetries. Similarly, to any 1d lattice system we attach a 2-group (modeled as a crossed module) and interpret the Nayak-Else anomaly index as an obstruction for promoting a group action to a morphism of 2-groups. The meaning of indices of Symmetry Protected Topological states is also illuminated by higher group symmetry.

math-ph↗

arXiv2025

AddressVLM: Cross-view Alignment Tuning for Image Address Localization using Large Vision-Language Models

Large visual language models (LVLMs) have demonstrated impressive performance in coarse-grained geo-localization at the country or city level, but they struggle with fine-grained street-level localization within urban areas. In this paper, we explore integrating city-wide address localization capabilities into LVLMs, facilitating flexible address-related question answering using street-view images. A key challenge is that the street-view visual question-and-answer (VQA) data provides only microscopic visual cues, leading to subpar performance in fine-tuned models. To tackle this issue, we incorporate perspective-invariant satellite images as macro cues and propose cross-view alignment tuning including a satellite-view and street-view image grafting mechanism, along with an automatic label generation mechanism. Then LVLM's global understanding of street distribution is enhanced through cross-view matching. Our proposed model, named AddressVLM, consists of two-stage training protocols: cross-view alignment tuning and address localization tuning. Furthermore, we have constructed two street-view VQA datasets based on image address localization datasets from Pittsburgh and San Francisco. Qualitative and quantitative evaluations demonstrate that AddressVLM outperforms counterpart LVLMs by over 9% and 12% in average address localization accuracy on these two datasets, respectively.

cs.CV↗

arXiv2024

AddressCLIP: Empowering Vision-Language Models for City-wide Image Address Localization

In this study, we introduce a new problem raised by social media and photojournalism, named Image Address Localization (IAL), which aims to predict the readable textual address where an image was taken. Existing two-stage approaches involve predicting geographical coordinates and converting them into human-readable addresses, which can lead to ambiguity and be resource-intensive. In contrast, we propose an end-to-end framework named AddressCLIP to solve the problem with more semantics, consisting of two key ingredients: i) image-text alignment to align images with addresses and scene captions by contrastive learning, and ii) image-geography matching to constrain image features with the spatial distance in terms of manifold learning. Additionally, we have built three datasets from Pittsburgh and San Francisco on different scales specifically for the IAL problem. Experiments demonstrate that our approach achieves compelling performance on the proposed datasets and outperforms representative transfer learning methods for vision-language models. Furthermore, extensive ablations and visualizations exhibit the effectiveness of the proposed method. The datasets and source code are available at https://github.com/xsx1001/AddressCLIP.

cs.CV↗

arXiv2024

Defying Imbalanced Forgetting in Class Incremental Learning

We observe a high level of imbalance in the accuracy of different classes in the same old task for the first time. This intriguing phenomenon, discovered in replay-based Class Incremental Learning (CIL), highlights the imbalanced forgetting of learned classes, as their accuracy is similar before the occurrence of catastrophic forgetting. This discovery remains previously unidentified due to the reliance on average incremental accuracy as the measurement for CIL, which assumes that the accuracy of classes within the same task is similar. However, this assumption is invalid in the face of catastrophic forgetting. Further empirical studies indicate that this imbalanced forgetting is caused by conflicts in representation between semantically similar old and new classes. These conflicts are rooted in the data imbalance present in replay-based CIL methods. Building on these insights, we propose CLass-Aware Disentanglement (CLAD) to predict the old classes that are more likely to be forgotten and enhance their accuracy. Importantly, CLAD can be seamlessly integrated into existing CIL methods. Extensive experiments demonstrate that CLAD consistently improves current replay-based methods, resulting in performance gains of up to 2.56%.

cs.CV↗

arXiv2023

Exploring Domain Incremental Video Highlights Detection with the LiveFood Benchmark

Video highlights detection (VHD) is an active research field in computer vision, aiming to locate the most user-appealing clips given raw video inputs. However, most VHD methods are based on the closed world assumption, i.e., a fixed number of highlight categories is defined in advance and all training data are available beforehand. Consequently, existing methods have poor scalability with respect to increasing highlight domains and training data. To address above issues, we propose a novel video highlights detection method named Global Prototype Encoding (GPE) to learn incrementally for adapting to new domains via parameterized prototypes. To facilitate this new research direction, we collect a finely annotated dataset termed LiveFood, including over 5,100 live gourmet videos that consist of four domains: ingredients, cooking, presentation, and eating. To the best of our knowledge, this is the first work to explore video highlights detection in the incremental learning setting, opening up new land to apply VHD for practical scenarios where both the concerned highlight domains and training data increase over time. We demonstrate the effectiveness of GPE through extensive experiments. Notably, GPE surpasses popular domain incremental learning methods on LiveFood, achieving significant mAP improvements on all domains. Concerning the classic datasets, GPE also yields comparable performance as previous arts. The code is available at: https://github.com/ForeverPs/IncrementalVHD_GPE.

cs.CV↗

arXiv2016

Customizable Precision of Floating-Point Arithmetic with Bitslice Vector Types

Customizing the precision of data can provide attractive trade-offs between accuracy and hardware resources. We propose a novel form of vector computing aimed at arrays of custom-precision floating point data. We represent these vectors in bitslice format. Bitwise instructions are used to implement arithmetic circuits in software that operate on customized bit-precision. Experiments show that this approach can be efficient for vectors of low-precision custom floating point types, while providing arbitrary bit precision.

cs.OH↗