Researcher profile

Shilong Liu

Shilong Liu contributes to research discovery and scholarly infrastructure.

ResearcherAffiliation not importedOpen to collaborate

Computer Vision Computation and Language physics.optics quant-ph Artificial Intelligence nlin.PS Information Retrieval physics.app-ph

Trust snapshot

Quick read

Trust 21 - EmergingVerification L1Unclaimed author

11works

0followers

8topics

4close collaborators

Actions

Decide how to stay connected

Follow researcher0

Claiming links this public author record to a researcher profile and unlocks direct collaboration workflows.

Direct collaboration

Open a focused conversation when the fit is right

Claim this author entity first to unlock direct invitations.

Inspect adjacent work, topics, institutions and collaborators without jumping out to a separate graph page.

Building this graph slice

BZPEER is loading the nearby papers, people, topics and institutions for this page.

preprint2026arXiv

A Survey of Self-Evolving Agents: What, When, How, and Where to Evolve on the Path to Artificial Super Intelligence

Large Language Models (LLMs) have demonstrated remarkable capabilities across diverse tasks but remain fundamentally static, unable to adapt their internal parameters to novel tasks, evolving knowledge domains, or dynamic interaction contexts. As LLMs are increasingly deployed in open-ended, interactive environments, this static nature has become a critical bottleneck, necessitating agents that can adaptively reason, act, and evolve in real time. This paradigm shift -- from scaling static models to developing self-evolving agents -- has sparked growing interest in architectures and methods enabling continual learning and adaptation from data, interactions, and experiences. This survey provides the first systematic and comprehensive review of self-evolving agents, organizing the field around three foundational dimensions: what, when, and how to evolve. We examine evolutionary mechanisms across agent components (e.g., models, memory, tools, architecture), categorize adaptation methods by stages (e.g., intra-test-time, inter-test-time), and analyze the algorithmic and architectural designs that guide evolutionary adaptation (e.g., scalar rewards, textual feedback, single-agent and multi-agent systems). Additionally, we analyze evaluation metrics and benchmarks tailored for self-evolving agents, highlight applications in domains such as coding, education, and healthcare, and identify critical challenges and research directions in safety, scalability, and co-evolutionary dynamics. By providing a structured framework for understanding and designing self-evolving agents, this survey establishes a roadmap for advancing more adaptive, robust, and versatile agentic systems in both research and real-world deployments, and ultimately sheds light on the realization of Artificial Super Intelligence (ASI) where agents evolve autonomously and perform beyond human-level intelligence across tasks.

preprint2026arXiv

CubeBench: Diagnosing Interactive, Long-Horizon Spatial Reasoning Under Partial Observations

Large Language Model (LLM) agents, while proficient in the digital realm, face a significant gap in physical-world deployment due to the challenge of forming and maintaining a robust spatial mental model. We identify three core cognitive challenges hindering this transition: spatial reasoning, long-horizon state tracking via mental simulation, and active exploration under partial observation. To isolate and evaluate these faculties, we introduce CubeBench, a novel generative benchmark centered on the Rubik's Cube. CubeBench uses a three-tiered diagnostic framework that progressively assesses agent capabilities, from foundational state tracking with full symbolic information to active exploration with only partial visual data. Our experiments on leading LLMs reveal critical limitations, including a uniform 0.00% pass rate on all long-horizon tasks, exposing a fundamental failure in long-term planning. We also propose a diagnostic framework to isolate these cognitive bottlenecks by providing external solver tools. By analyzing the failure modes, we provide key insights to guide the development of more physically-grounded intelligent agents.

preprint2026arXiv

Learning Agent Routing From Early Experience

LLM agents achieve strong performance on complex reasoning tasks but incur high latency and compute cost. In practice, many queries fall within the capability boundary of cutting-edge LLMs and do not require full agent execution, making effective routing between LLMs and agents a key challenge. We study the problem of routing queries between lightweight LLM inference and full agent execution under realistic cold-start settings. To address this, we propose BoundaryRouter, a training-free routing framework that uses early behavioral experience and rubric-guided reasoning to decide whether to answer a query with direct LLM inference or escalate to an agent. BoundaryRouter builds a compact experience memory by executing both systems on a shared seed set and retrieves similar cases at inference time to guide routing decisions. To evaluate this method, we introduce RouteBench, a benchmark covering in-domain, paraphrased, and out-of-domain route settings. Experiments show that BoundaryRouter reduces inference time by 60.6% compared to the agent while improving performance by 28.6% over direct LLM inference, outperforming prompt-based and retrieval-only routing by an average of 37.9% and 8.2%, respectively.

preprint2026arXiv

MemEye: A Visual-Centric Evaluation Framework for Multimodal Agent Memory

Long-term agent memory is increasingly multimodal, yet existing evaluations rarely test whether agents preserve the visual evidence needed for later reasoning. In prior work, many visually grounded questions can be answered using only captions or textual traces, allowing answers to be inferred without preserving the fine-grained visual evidence. Meanwhile, harder cases that require reasoning over changing visual states are largely absent. Therefore, we introduce MemEye, a framework that evaluates memory capabilities from two dimensions: one measures the granularity of decisive visual evidence (from scene-level to pixel-level evidence), and the other measures how retrieved evidence must be used (from single evidence to evolutionary synthesis). Under this framework, we construct a new benchmark across 8 life-scenario tasks, with ablation-driven validation gates for assessing answerability, shortcut resistance, visual necessity, and reasoning structure. By evaluating 13 memory methods across 4 VLM backbones, we show that current architectures still struggle to preserve fine-grained visual details and reason about state changes over time. Our findings show that long-term multimodal memory depends on evidence routing, temporal tracking, and detail extraction.

preprint2022arXiv

DAB-DETR: Dynamic Anchor Boxes are Better Queries for DETR

We present in this paper a novel query formulation using dynamic anchor boxes for DETR (DEtection TRansformer) and offer a deeper understanding of the role of queries in DETR. This new formulation directly uses box coordinates as queries in Transformer decoders and dynamically updates them layer-by-layer. Using box coordinates not only helps using explicit positional priors to improve the query-to-feature similarity and eliminate the slow training convergence issue in DETR, but also allows us to modulate the positional attention map using the box width and height information. Such a design makes it clear that queries in DETR can be implemented as performing soft ROI pooling layer-by-layer in a cascade manner. As a result, it leads to the best performance on MS-COCO benchmark among the DETR-like detection models under the same setting, e.g., AP 45.7\% using ResNet50-DC5 as backbone trained in 50 epochs. We also conducted extensive experiments to confirm our analysis and verify the effectiveness of our methods. Code is available at \url{https://github.com/SlongLiu/DAB-DETR}.

preprint2022arXiv

DINO: DETR with Improved DeNoising Anchor Boxes for End-to-End Object Detection

We present DINO (\textbf{D}ETR with \textbf{I}mproved de\textbf{N}oising anch\textbf{O}r boxes), a state-of-the-art end-to-end object detector. % in this paper. DINO improves over previous DETR-like models in performance and efficiency by using a contrastive way for denoising training, a mixed query selection method for anchor initialization, and a look forward twice scheme for box prediction. DINO achieves $49.4$AP in $12$ epochs and $51.3$AP in $24$ epochs on COCO with a ResNet-50 backbone and multi-scale features, yielding a significant improvement of $\textbf{+6.0}$\textbf{AP} and $\textbf{+2.7}$\textbf{AP}, respectively, compared to DN-DETR, the previous best DETR-like model. DINO scales well in both model size and data size. Without bells and whistles, after pre-training on the Objects365 dataset with a SwinL backbone, DINO obtains the best results on both COCO \texttt{val2017} ($\textbf{63.2}$\textbf{AP}) and \texttt{test-dev} (\textbf{$\textbf{63.3}$AP}). Compared to other models on the leaderboard, DINO significantly reduces its model size and pre-training data size while achieving better results. Our code will be available at \url{https://github.com/IDEACVR/DINO}.

preprint2022arXiv

Experimental realisations of the fractional Schrödinger equation in the temporal domain

The fractional Schrödinger equation (FSE) -- a natural extension of the standard Schrödinger equation -- is the basis of fractional quantum mechanics. It can be obtained by replacing the kinetic-energy operator with a fractional derivative. Here, we report the experimental realisation of an optical FSE for femtosecond laser pulses in the temporal domain. Programmable holograms and the single-shot measurement technique are respectively used to emulate a \textit{Lévy waveguide} and to reconstruct the amplitude and phase of the pulses. Varying the Lévy index of the FSE and the initial pulse, the temporal dynamics is observed in diverse forms, including solitary, splitting and merging pulses, double Airy modes, and ``rain-like'' multi-pulse patterns. Furthermore, the transmission of input pulses carrying a fractional phase exhibits a ``fractional-phase protection'' effect through a regular (non-fractional) material. The experimentally generated fractional time-domain pulses offer the potential for designing optical signal-processing schemes.

preprint2022arXiv

On-demand harnessing of photonic soliton molecules

Soliton molecules (SMs) are fundamentally important modes in nonlinear optical systems. It is a challenge to experimentally produce SMs with a required temporal separation in mode-locked fiber lasers. Here, we propose and realize an experimental scenario for harnessing SM dynamics in a laser setup. In particular, we tailor SMs in a mode-locked laser controlled by second-order group-velocity dispersion and dispersion losses: the real part of dispersion maintains the balance between the dispersion and nonlinearity, while the dispersion loss determines the balance of gain and losses. The experimental results demonstrate that the dispersion loss makes it possible to select desired values of the temporal separation (TS) in bound pairs of SMs in the system. Tunability of the SM's central wavelength and the corresponding hysteresis are addressed too. The demonstrated regime allows us to create multiple SMs with preselected values of the TS and central wavelength, which shows the potential of our setup for the design of optical data-processing schemes.

preprint2022arXiv

Vision-Language Intelligence: Tasks, Representation Learning, and Large Models

This paper presents a comprehensive survey of vision-language (VL) intelligence from the perspective of time. This survey is inspired by the remarkable progress in both computer vision and natural language processing, and recent trends shifting from single modality processing to multiple modality comprehension. We summarize the development in this field into three time periods, namely task-specific methods, vision-language pre-training (VLP) methods, and larger models empowered by large-scale weakly-labeled data. We first take some common VL tasks as examples to introduce the development of task-specific methods. Then we focus on VLP methods and comprehensively review key components of the model structures and training methods. After that, we show how recent work utilizes large-scale raw image-text data to learn language-aligned visual representations that generalize better on zero or few shot learning tasks. Finally, we discuss some potential future trends towards modality cooperation, unified representation, and knowledge incorporation. We believe that this review will be of help for researchers and practitioners of AI and ML, especially those interested in computer vision and natural language processing.

preprint2020arXiv

Increasing two-photon entangled dimensions by shaping input beam profiles

Photon pair entangled in high dimensional orbital angular momentum (OAM) degree of freedom (DOF) has been widely regarded as a possible source in improving the capacity of quantum information processing. The need for the generation of a high dimensional maximally entangled state in the OAM DOF is therefore much desired. In this work, we demonstrate a simple method to generate a broader and flatter OAM spectrum, i.e. a larger spiral bandwidth (SB), of entangled photon pairs generated through spontaneous parametric down-conversion by modifying the pump beam profile. By investigating both experimentally and theoretically, we have found that an exponential pump profile that is roughly the inverse of the mode profiles of the single-mode fibers used for OAM detection will provide a much larger SB when compared to a Gaussian shaped pump.

preprint2019arXiv

A high-dimensional quantum frequency converter

In high dimensional quantum communication networks, quantum frequency convertor (QFC) is indispensable as an interface in the frequency domain. For example, many QFCs have been built to link atomic memories and fiber channels. However, almost all of QFCs work in a two-dimensional space. It is still a pivotal challenge to construct a high-quality QFC for some complex quantum states, e.g., a high dimensional single-photon state that refers to a qudit. Here, we firstly propose a high-dimensional QFC for an orbital angular momentum qudit via sum frequency conversion with a flat top beam pump. As a proof-of-principle demonstration, we realize quantum frequency conversions for a qudit from infrared to visible range. Based on the qudit quantum state tomography, the fidelities of converted state are 98.29(95.02)\%, 97.42(91.74)\%, and 86.75(67.04)\% for a qudit without (with) dark counts in 2,3, and 5 dimensions, respectively. The demonstration is very promising for constructing a high capacity quantum communication network.

Institution

Affiliation not imported yet

This author record came from a source that does not expose affiliation metadata. Once the author claims the profile or we enrich the record from another provider, this section will link to the concrete institution.

Topic footprint