Searcharxiv⌕ Search

arXiv subjects

Chi Li

Publications and source records attributed to Chi Li.

At least 19 recordsLinked to original sources
arXiv2026

Quantum skyrmion parallelism via metasurface-tailored high-dimensional entanglement

Quantum skyrmions are topological structures that have garnered significant interest due to their demonstrated robustness and versatility across diverse optical platforms. However, existing approaches for their generation are limited to producing pre-determined two dimensional qubit states with a single topology. Here we create multi-dimensional topological states by introducing a non-local interaction between high-dimensional photonic entanglement and a metasurface, where the topological transformation induced by the metasurface is made non-deterministic by the probabilistic nature of the interfacing entangled state. Within this framework, we demonstrate that individual quantum states can host multiple co-existing topologies that are only revealed upon measurement, allowing for their parallel transport within distinct spatial mode channels. We confirm this by revealing the rich topological landscape within our modified Hilbert space while controlling the desired output topology by orbital angular momentum (OAM) projections on one of the entangled photons, producing multiple non-local polarization-OAM entangled states characterized by distinct topological classes, all from a single metasurface device. Our results reveal new capability when structured high-dimensional entanglement is interfaced with structured matter capable of coupling photonic degrees of freedom, establishing a new pathway for the compact generation of complex quantum states.

quant-ph↗
arXiv2026

Metalens stereoscopic optical palpation for imaging cancer mechanics

Accurate intraoperative margin assessment is critical for complete resection of solid tumours. However, surgeons routinely rely on manual palpation, which is highly subjective and fundamentally limited by poor spatial resolution, and pre-operative imaging lacks real-time feedback. While emerging optical elastography techniques offer quantitative mechanical contrast, their reliance on bulky optical components necessitates cumbersome alignment and hinders integration into miniaturised surgical instruments. Here, we report metalens stereoscopic optical palpation (MSOP), a nanophotonic imaging platform leveraging a binocular metalens integrated with a single CMOS sensor to achieve high-fidelity, self-aligned elastography. By replacing complex bulk optics with a compact stereoscopic architecture, MSOP enables robust, high-contrast mapping of tissue surface stress. Following validation in heterogeneous silicone phantoms, we demonstrate MSOP's translational potential by characterising the mechanical signatures of malignancies in fresh mouse pancreatic cancer models, as well as excised human breast and liver specimens. By providing reliable, label-free mechanical contrast across diverse oncological landscapes, MSOP offers a compact tool for precise intraoperative margin delineation to reduce re-excision rates and improve surgical outcomes.

physics.optics↗
arXiv2026

CoAnchor: Robust Collaborative Perception under Spatio-Temporal Misalignment via Object-Level Anchors

Collaborative perception extends the sensing range of a single vehicle by fusing observations from nearby agents, which improves the robustness of autonomous driving. In realistic deployments, however, the received collaborator messages are often affected by both communication delay and relative-pose noise, which jointly cause stale observations, spatial misalignment, and unstable feature fusion. Existing methods usually address these issues from either the spatial or temporal side, but handling them jointly in a unified and efficient manner remains challenging. In this paper, we propose CoAnchor, an anchor-centric spatio-temporal alignment framework for asynchronous collaborative perception. Instead of directly reasoning on dense BEV features, CoAnchor builds sparse object-level spatio-temporal anchors as a shared interface for pose correction and tightly connects spatial refinement, temporal propagation, and current-time verification within one unified loop, while keeping the overall correction process lightweight. Extensive experiments on both simulated and real-world datasets illustrate that CoAnchor remains competitive under clean settings and improves the robustness under joint delay and pose perturbations with a favorable practical accuracy-efficiency trade-off.

cs.CV↗
arXiv2026

FM-Bench: A Benchmark for Long-Horizon Management with Competing Agents

Language model agents now execute bounded tasks reliably. Whether they can sustain effective decision-making over long horizons, where actions have cumulative consequences and the environment responds to their choices, remains largely unmeasured. FM-Bench (Football Management Benchmark) measures this. An LLM agent runs a football club for 20 in-game years through 26 tools and roughly 340 to 400 decision stops. It drafts a squad on the same budget as every rival, trades players, negotiates contracts, invests in facilities and youth, sets lineups, and answers to a board that can fire it, while a deterministic engine accumulates every year into one final score with no LLM judge or human rater. The solo track plays each of 15 frontier models against a frozen scripted world, and the Arena places the same models plus a scripted anchor in one shared 20-year world; to our knowledge, the first head-to-head evaluation at this scale. We measure six behavioral capabilities behind the score. Across three seeds, all 15 models complete every horizon while the blind scripted baselines die out in most of theirs, and claude-fable-5 tops the solo board on mean score and the Arena, where the title nonetheless rotates among ten models. Neither scale, price, nor vendor predicts the order; the order settles only late in the horizon, and the best first-play human lands only at the bottom of the model board. What separates the models is managerial behavior rather than computation. Higher-scoring models reduce slow-payoff investment near the end, keep cash invested rather than idle, and open renewals well before the deadline, while token spend predicts nothing. No model learns the market's hidden prices from hundreds of rejected bids, and self-managed memory fails in two opposite modes: an archive that only grows or a plan rewritten every season. Code is available at https://github.com/Analogy-AI/fm-bench.

cs.AI↗
arXiv2026

The sharp volume gap for K\"ahler manifolds with positive Ricci curvature

We prove a sharp volume gap estimate: if an $n$-dimensional compact K\"ahler manifold $(X, \omega)$ satisfies $\mathrm{Ric}(\omega)\ge (n+1)\omega$ and $X\not\cong \mathbb{P}^n$, then $\mathrm{vol}(X, \omega)\le \frac{2n^n}{(n+1)^n}\mathrm{vol}(\mathbb{P}^n,\omega_{\mathrm{FS}})=\frac{2^{n+1} \, \pi^n \, n^n}{(n+1)^n}$. Moreover $\mathrm{vol}(X, \omega)= \frac{2^{n+1} \, \pi^n \, n^n}{(n+1)^n}$ occurs if and only if $(X, \omega)$ is biholomorphically isometric to the K\"ahler-Einstein metric on the quadric hypersurface $Q^n$ or on the product $\mathbb{P}^1\times \mathbb{P}^{n-1}$. We also obtain sharp volume gap estimates for K-semistable toric log Fano pairs.

math.DG↗
arXiv2026

FUPareto: Bridging the Forgetting-Utility Gap in Federated Unlearning via Pareto Augmented Optimization

Federated Unlearning (FU) aims to efficiently remove the influence of specific client data from a federated model while preserving utility for the remaining clients. However, three key challenges remain: (1) existing unlearning objectives often compromise model utility or increase vulnerability to Membership Inference Attacks (MIA); (2) there is a persistent conflict between forgetting and utility, where further unlearning inevitably harms retained performance; and (3) support for concurrent multi-client unlearning is poor, as gradient conflicts among clients degrade the quality of forgetting. To address these issues, we propose FUPareto, an efficient unlearning framework via Pareto-augmented optimization. We first introduce the Minimum Boundary Shift (MBS) Loss, which enforces unlearning by suppressing the target class logit below the highest non-target class logit; this can improve the unlearning efficiency and mitigate MIA risks. During the unlearning process, FUPareto performs Pareto improvement steps to preserve model utility and executes Pareto expansion to guarantee forgetting. Specifically, during Pareto expansion, the framework integrates a Null-Space Projected Multiple Gradient Descent Algorithm (MGDA) to decouple gradient conflicts. This enables effective, fair, and concurrent unlearning for multiple clients while minimizing utility degradation. Extensive experiments across diverse scenarios demonstrate that FUPareto consistently outperforms state-of-the-art FU methods in both unlearning efficacy and retained utility.

cs.LG↗
arXiv2025

RecGPT Technical Report

Recommender systems are among the most impactful applications of artificial intelligence, serving as critical infrastructure connecting users, merchants, and platforms. However, most current industrial systems remain heavily reliant on historical co-occurrence patterns and log-fitting objectives, i.e., optimizing for past user interactions without explicitly modeling user intent. This log-fitting approach often leads to overfitting to narrow historical preferences, failing to capture users' evolving and latent interests. As a result, it reinforces filter bubbles and long-tail phenomena, ultimately harming user experience and threatening the sustainability of the whole recommendation ecosystem. To address these challenges, we rethink the overall design paradigm of recommender systems and propose RecGPT, a next-generation framework that places user intent at the center of the recommendation pipeline. By integrating large language models (LLMs) into key stages of user interest mining, item retrieval, and explanation generation, RecGPT transforms log-fitting recommendation into an intent-centric process. To effectively align general-purpose LLMs to the above domain-specific recommendation tasks at scale, RecGPT incorporates a multi-stage training paradigm, which integrates reasoning-enhanced pre-alignment and self-training evolution, guided by a Human-LLM cooperative judge system. Currently, RecGPT has been fully deployed on the Taobao App. Online experiments demonstrate that RecGPT achieves consistent performance gains across stakeholders: users benefit from increased content diversity and satisfaction, merchants and the platform gain greater exposure and conversions. These comprehensive improvement results across all stakeholders validates that LLM-driven, intent-centric design can foster a more sustainable and mutually beneficial recommendation ecosystem.

cs.IR↗
arXiv2025

Disorder-enabled Synthetic Metasurfaces

Optical metasurfaces have catalyzed transformative advances across imaging, optoelectronics, quantum information processing, sensing, energy conversion, and optical computing. Yet, despite this rapid progress, most research remains focused on optimizing single functionalities, constrained by the persistent challenge of integrating multiple functions within a single device. Here, we demonstrate that engineered structural disorder of metapixels, used to implement a photonic function, can significantly reduce the area required across the entire aperture without compromising optical performance. The unallocated space can then be repurposed to encode functionally distinct metapixels without increasing the design complexity, each independently addressable via various optical degrees of freedom. As a proof of concept, we present a synthetic achromatic metalens featuring 11 spectrally distinct lens profiles encoded through nonlocal metapixels engineered to support sharp resonances via quasi bound states in the continuum. This large-scale metalens with 8.1 mm aperture achieves diffraction-limited achromatic focusing across the 1200 to 1400 nm spectral window. We further incorporate polarization-selective metapixels to implement momentum-space distinct gratings, enabling single-shot, high spatial resolution polarimetric imaging of arbitrarily structured light fields, including radial and azimuthal vector beams and optical skyrmions. Altogether, this disorder-enabled synthetic metasurface platform establishes a versatile foundation for unifying diverse photonic functionalities within a single optical element, marking a substantial step toward compact, high-density, multifunctional optical devices.

physics.optics↗
arXiv2025

On the volume of K-semistable Fano manifolds

We prove that the anti-canonical volume of an $n$-dimensional K-semistable Fano manifold that is not $\mathbb{P}^n$ is at most $2n^n$. Moreover, the volume is equal to $2n^n$ if and only if $X\cong \mathbb{P}^1\times \mathbb{P}^{n-1}$ or $X$ is a smooth quadric hypersurface $Q\subset \mathbb{P}^{n+1}$. Our proof is based on a new connection between K-semistability and minimal rational curves.

math.AG↗
arXiv2025

Resonant helical dichroism in twisted dielectric metastructures

Circular dichroism, arising from interactions with light fields of opposite spin angular momentum, has become a fundamental tool for molecular characterization. Meanwhile, helical dichroism (HD) - the dichroic response to vortex beams carrying opposite orbital angular momentum (OAM) - offers an alternative approach for probing chiral molecules and photonic structures. Previous demonstrations of HD have been limited to non-resonant light-matter interactions with chiral micro- and nanostructures, leaving the realization of resonance helical dichroism largely unexplored. Here, we present the design and implementation of twisted dielectric metastructures, composed of an array of rotated silicon trimer nanostructures harnessing nonlocal photonic modes with a high quality factor of several dozen that enable strong resonant HD for OAM values up to $10$. We experimentally demonstrate resonantly enhanced HD for strongly focused OAM beams with the magnitude of topological charges from $1$ to $3$. Our findings pave the way for resonant nanophotonics involving OAM beams, unlocking the full potential of structured light for applications in molecular sensing, optical imaging, nonlinear optics, and optical data storage.

physics.optics↗
arXiv2025

Valley optoelectronics based on meta-waveguide photodetectors

In transition metal dichalcogenides, the valley degree of freedom directly couples valley-polarised excitons - excited by circularly polarised light - to valley-dependent chiral photons, enabling ultrafast light-driven valleytronics. However, achieving fully integrated valley optoelectronics - incorporating on-chip generation, selective routing, and electrical readout of valley-dependent chiral photons - remains an unresolved challenge. We present a valley-driven hybrid nanophotonic-optoelectronic circuit that integrates chirality-selective meta-waveguide photodetectors with transition metal dichalcogenides. At room temperature, our purposely designed meta-waveguide device generates near-unity valley-dependent chiral photons in the second harmonic generation from an encapsulated tungsten disulfide monolayer and selectively couples them to unidirectional waveguide modes, achieving an exceptional polarisation selectivity of 0.97. These valley-dependent waveguide modes were subsequently detected by atomically thin few-layer tungsten diselenide photodetectors, exclusively responsive to the above-bandgap upconverted photons, thereby enabling all-on-chip processing of valley-multiplexed images. Our demonstration bridges a critical gap in lightwave valleytronics, paving the way for compact, scalable valley information processing and fostering the development of light-based valleytronic quantum technologies.

physics.optics↗
arXiv2025

Minimal log discrepancy and orbifold curves

We show that the minimal log discrepancy of any isolated Fano cone singularity is at most the dimension of the variety. This is based on its relation with dimensions of moduli spaces of orbifold rational curves. We also propose a conjectural characterization of weighted projective spaces as Fano orbifolds in terms of orbifold rational curves, which would imply the equality holds only for smooth points.

math.AG↗
arXiv2025

An Inverse Design Wavelength Demultiplexer for On-Chip Photoluminescence Sorting in TMDC Heterostructures

Emerging two-dimensional transition metal dichalcogenides (TMDCs) offer a promising platform for on-chip integrated photonics because of their unique optical and electronic properties. Their naturally passivated surfaces make them highly tolerant to lattice mismatch, enabling seamless heterogeneous integration by stacking different van der Waals materials, a crucial step in the development of advanced photonic devices. Here, we demonstrate the use of an inverse design wavelength demultiplexing waveguides for on-chip sorting and routing of distinct photoluminescence from the heterojunction formed by WS2 and WSe2 monolayers. The integrated nanophotonic chip splits and sorts excitonic emission into individual waveguides at room temperature. Our demonstration opens up new perspectives for integrating light sources in van der Waals materials with functional integrated photonics, offering a versatile platform for both fundamental research and practical applications.

physics.optics↗
arXiv2024

Multi-Objective Large Language Model Unlearning

Machine unlearning in the domain of large language models (LLMs) has attracted great attention recently, which aims to effectively eliminate undesirable behaviors from LLMs without full retraining from scratch. In this paper, we explore the Gradient Ascent (GA) approach in LLM unlearning, which is a proactive way to decrease the prediction probability of the model on the target data in order to remove their influence. We analyze two challenges that render the process impractical: gradient explosion and catastrophic forgetting. To address these issues, we propose Multi-Objective Large Language Model Unlearning (MOLLM) algorithm. We first formulate LLM unlearning as a multi-objective optimization problem, in which the cross-entropy loss is modified to the unlearning version to overcome the gradient explosion issue. A common descent update direction is then calculated, which enables the model to forget the target data while preserving the utility of the LLM. Our empirical results verify that MoLLM outperforms the SOTA GA-based LLM unlearning methods in terms of unlearning effect and model utility preservation. The source code is available at https://github.com/zibinpan/MOLLM.

cs.CL↗
arXiv2024

Federated Unlearning with Gradient Descent and Conflict Mitigation

Federated Learning (FL) has received much attention in recent years. However, although clients are not required to share their data in FL, the global model itself can implicitly remember clients' local data. Therefore, it's necessary to effectively remove the target client's data from the FL global model to ease the risk of privacy leakage and implement ``the right to be forgotten". Federated Unlearning (FU) has been considered a promising way to remove data without full retraining. But the model utility easily suffers significant reduction during unlearning due to the gradient conflicts. Furthermore, when conducting the post-training to recover the model utility, the model is prone to move back and revert what has already been unlearned. To address these issues, we propose Federated Unlearning with Orthogonal Steepest Descent (FedOSD). We first design an unlearning Cross-Entropy loss to overcome the convergence issue of the gradient ascent. A steepest descent direction for unlearning is then calculated in the condition of being non-conflicting with other clients' gradients and closest to the target client's gradient. This benefits to efficiently unlearn and mitigate the model utility reduction. After unlearning, we recover the model utility by maintaining the achievement of unlearning. Finally, extensive experiments in several FL scenarios verify that FedOSD outperforms the SOTA FU algorithms in terms of unlearning and model utility.

cs.LG↗
arXiv2024

Ultrathin BIC metasurfaces based on ultralow-loss Sb2Se3 phase-change material

Phase-change materials (PCMs) are increasingly recognised as promising platforms for tunable photonic devices due to their ability to modulate optical properties through solid-state phase transitions. Ultrathin and low-loss PCMs are highly valued for their fast and more effective phase transitions and applications in reconfigurable photonic chips, metasurfaces, optical modulators, sensors, photonic memories, and neuromorphic computing. However, conventional PCMs such as GST, GSST, VO2, and In3SbTe2, despite optimisation for tunable meta-optics, suffer from high intrinsic losses in the near-infrared (NIR) region, limiting their potential for high quality factor (Q-factor) resonant metasurfaces. Here we present the design and fabrication of tunable bound states in the continuum (BIC) metasurfaces using the ultralow-loss PCM Sb2Se3. Our BIC metasurfaces, only 25 nm thick, achieve high modulation depth and broad resonance tuning in the NIR with high Q-factors up to 130, without the need for additional materials. Experimentally, we employ these BIC metasurfaces to modulate photoluminescence in rare earth-doped upconversion nanoparticles, reducing the excitation power for multiphoton photoluminescence and enabling emission polarisation manipulation. This work offers a promising platform for developing active resonant metasurfaces in the NIR region, with broad applications including super resolution imaging, optical modulation, ultrafast switches, harmonic generation, colour filtering, and optical sensing.

physics.optics↗
arXiv2024

K\"{a}hler compactification of $\mathbb{C}^n$ and Reeb dynamics

Let $X$ be a smooth complex manifold. Assume that $Y\subset X$ is a K\"{a}hler submanifold such that $X\setminus Y$ is biholomorphic to $\mathbb{C}^n$. We prove that $(X, Y)$ is biholomorphic to the standard example $(\mathbb{P}^n, \mathbb{P}^{n-1})$. We then study certain K\"{a}hler orbifold compactifications of $\mathbb{C}^n$ and, as an application, prove that on $\mathbb{C}^3$ the flat metric is the only asymptotically conical Ricci-flat K\"{a}hler metric whose metric cone at infinity has a smooth link. As a key technical ingredient, we derive a new characterization of minimal discrepancy of isolated Fano cone singularities by using $S^1$-equivariant positive symplectic homology.

math.DG↗