arXiv subjects

Yifan Liu

Publications and source records attributed to Yifan Liu.

At least 127 records · Page 7Linked to original sources

arXiv2024

AlignRec: Aligning and Training in Multimodal Recommendations

With the development of multimedia systems, multimodal recommendations are playing an essential role, as they can leverage rich contexts beyond interactions. Existing methods mainly regard multimodal information as an auxiliary, using them to help learn ID features; However, there exist semantic gaps among multimodal content features and ID-based features, for which directly using multimodal information as an auxiliary would lead to misalignment in representations of users and items. In this paper, we first systematically investigate the misalignment issue in multimodal recommendations, and propose a solution named AlignRec. In AlignRec, the recommendation objective is decomposed into three alignments, namely alignment within contents, alignment between content and categorical ID, and alignment between users and items. Each alignment is characterized by a specific objective function and is integrated into our multimodal recommendation framework. To effectively train AlignRec, we propose starting from pre-training the first alignment to obtain unified multimodal features and subsequently training the following two alignments together with these features as input. As it is essential to analyze whether each multimodal feature helps in training and accelerate the iteration cycle of recommendation models, we design three new classes of metrics to evaluate intermediate performance. Our extensive experiments on three real-world datasets consistently verify the superiority of AlignRec compared to nine baselines. We also find that the multimodal features generated by AlignRec are better than currently used ones, which are to be open-sourced in our repository https://github.com/sjtulyf123/AlignRec_CIKM24.

cs.IR↗

arXiv2024

GTP-4o: Modality-prompted Heterogeneous Graph Learning for Omni-modal Biomedical Representation

Recent advances in learning multi-modal representation have witnessed the success in biomedical domains. While established techniques enable handling multi-modal information, the challenges are posed when extended to various clinical modalities and practical modalitymissing setting due to the inherent modality gaps. To tackle these, we propose an innovative Modality-prompted Heterogeneous Graph for Omnimodal Learning (GTP-4o), which embeds the numerous disparate clinical modalities into a unified representation, completes the deficient embedding of missing modality and reformulates the cross-modal learning with a graph-based aggregation. Specially, we establish a heterogeneous graph embedding to explicitly capture the diverse semantic properties on both the modality-specific features (nodes) and the cross-modal relations (edges). Then, we design a modality-prompted completion that enables completing the inadequate graph representation of missing modality through a graph prompting mechanism, which generates hallucination graphic topologies to steer the missing embedding towards the intact representation. Through the completed graph, we meticulously develop a knowledge-guided hierarchical cross-modal aggregation consisting of a global meta-path neighbouring to uncover the potential heterogeneous neighbors along the pathways driven by domain knowledge, and a local multi-relation aggregation module for the comprehensive cross-modal interaction across various heterogeneous relations. We assess the efficacy of our methodology on rigorous benchmarking experiments against prior state-of-the-arts. In a nutshell, GTP-4o presents an initial foray into the intriguing realm of embedding, relating and perceiving the heterogeneous patterns from various clinical modalities holistically via a graph theory. Project page: https://gtp-4-o.github.io/.

cs.CV↗

arXiv2024

EndoSparse: Real-Time Sparse View Synthesis of Endoscopic Scenes using Gaussian Splatting

3D reconstruction of biological tissues from a collection of endoscopic images is a key to unlock various important downstream surgical applications with 3D capabilities. Existing methods employ various advanced neural rendering techniques for photorealistic view synthesis, but they often struggle to recover accurate 3D representations when only sparse observations are available, which is usually the case in real-world clinical scenarios. To tackle this {sparsity} challenge, we propose a framework leveraging the prior knowledge from multiple foundation models during the reconstruction process, dubbed as \textit{EndoSparse}. Experimental results indicate that our proposed strategy significantly improves the geometric and appearance quality under challenging sparse-view conditions, including using only three views. In rigorous benchmarking experiments against state-of-the-art methods, \textit{EndoSparse} achieves superior results in terms of accurate geometry, realistic appearance, and rendering efficiency, confirming the robustness to sparse-view limitations in endoscopic reconstruction. \textit{EndoSparse} signifies a steady step towards the practical deployment of neural 3D reconstruction in real-world clinical scenarios. Project page: https://endo-sparse.github.io/.

cs.CV↗

arXiv2024

GaussianStego: A Generalizable Stenography Pipeline for Generative 3D Gaussians Splatting

Recent advancements in large generative models and real-time neural rendering using point-based techniques pave the way for a future of widespread visual data distribution through sharing synthesized 3D assets. However, while standardized methods for embedding proprietary or copyright information, either overtly or subtly, exist for conventional visual content such as images and videos, this issue remains unexplored for emerging generative 3D formats like Gaussian Splatting. We present GaussianStego, a method for embedding steganographic information in the rendering of generated 3D assets. Our approach employs an optimization framework that enables the accurate extraction of hidden information from images rendered using Gaussian assets derived from large models, while maintaining their original visual quality. We conduct preliminary evaluations of our method across several potential deployment scenarios and discuss issues identified through analysis. GaussianStego represents an initial exploration into the novel challenge of embedding customizable, imperceptible, and recoverable information within the renders produced by current 3D generative models, while ensuring minimal impact on the rendered content's quality.

cs.CV↗

arXiv2024

Low Excess Noise, High Quantum Efficiency Avalanche Photodiodes for Beyond 2 μm Wavelength Detection

The increasing concentration of greenhouse gases, notably CH4 and CO2, has fueled global temperature increases, intensifying concerns regarding the prevailing climate crisis. Effectively monitoring these gases demands a detector spanning the extended short-wavelength infrared (~2.4 μm) range, covering wavelengths of CH4 (1.65 μm) and CO2 (2.05 μm). The state-of-the-art HgCdTe avalanche photodetectors (APDs) offer exceptional performance metrics, including high gain (M) and low excess noise (F). However, their widespread adoption is hindered by inherent challenges such as manufacturability, reproducibility, and cost factors. Moreover, their reliance on cryogenic cooling adds to the cost, size, weight, and power of the system. We have demonstrated a linear mode APD combining an InGaAs/GaAsSb type-II superlattice absorber and an AlGaAsSb multiplier lattice matched to InP substrates. This APD has demonstrated a room temperature M of 178, a maximum measurable external quantum efficiency of 3560 % at 2 μm, an extremely low excess noise (F < 2 at M < 20), and a small temperature coefficient of breakdown (7.58 mV/K μm). Such a high performance APD with manufacturable semiconductor materials could lead to a rapid transition to a commercial III-V foundry, holding the promise of revolutionizing high-sensitivity receivers for greenhouse gas monitoring.

cond-mat.mtrl-sci↗

arXiv2024

LGS: A Light-weight 4D Gaussian Splatting for Efficient Surgical Scene Reconstruction

The advent of 3D Gaussian Splatting (3D-GS) techniques and their dynamic scene modeling variants, 4D-GS, offers promising prospects for real-time rendering of dynamic surgical scenarios. However, the prerequisite for modeling dynamic scenes by a large number of Gaussian units, the high-dimensional Gaussian attributes and the high-resolution deformation fields, all lead to serve storage issues that hinder real-time rendering in resource-limited surgical equipment. To surmount these limitations, we introduce a Lightweight 4D Gaussian Splatting framework (LGS) that can liberate the efficiency bottlenecks of both rendering and storage for dynamic endoscopic reconstruction. Specifically, to minimize the redundancy of Gaussian quantities, we propose Deformation-Aware Pruning by gauging the impact of each Gaussian on deformation. Concurrently, to reduce the redundancy of Gaussian attributes, we simplify the representation of textures and lighting in non-crucial areas by pruning the dimensions of Gaussian attributes. We further resolve the feature field redundancy caused by the high resolution of 4D neural spatiotemporal encoder for modeling dynamic scenes via a 4D feature field condensation. Experiments on public benchmarks demonstrate efficacy of LGS in terms of a compression rate exceeding 9 times while maintaining the pleasing visual quality and real-time rendering efficiency. LGS confirms a substantial step towards its application in robotic surgical services.

cs.CV↗

arXiv2024

Ultrastable vacuum-gap Fabry-Pérot cavities operated in air

We demonstrate a vacuum-gap ultrastable optical reference cavity that does not require a vacuum enclosure. Our simple method of optical contact bonding in a vacuum environment allows for cavity operation in air while maintaining vacuum between the cavity mirrors. Vacuum is maintained long term, with no observed degradation in cavity stability for over 1 year after bonding. For a 1550 nm laser stabilized to a 9.7 mL in-vacuum bonded cavity, the measured Allan deviation is $2.4\times 10^{-14}$ at 1 s and its phase noise is thermal-noise-limited from 0.1 Hz to 10 kHz, reaching about -105 dBc/Hz at 10 kHz offset frequency. This represents the highest stability of any oscillator operated without a vacuum enclosure. Furthermore, we demonstrate a 0.5 mL in-vacuum bonded cavity created using microfabricated mirrors and cavity dicing, with phase noise reaching -95 dBc/Hz at 10 kHz offset frequency. By relieving the need for high-vacuum enclosures, we greatly enhance the portability and utility of low noise, compact cavity-stabilized lasers, with applications ranging from environmental sensing to mobile optical clocks to ultralow noise microwave generation.

physics.optics↗

arXiv2024

Retrieval Augmented Fact Verification by Synthesizing Contrastive Arguments

The rapid propagation of misinformation poses substantial risks to public interest. To combat misinformation, large language models (LLMs) are adapted to automatically verify claim credibility. Nevertheless, existing methods heavily rely on the embedded knowledge within LLMs and / or black-box APIs for evidence collection, leading to subpar performance with smaller LLMs or upon unreliable context. In this paper, we propose retrieval augmented fact verification through the synthesis of contrasting arguments (RAFTS). Upon input claims, RAFTS starts with evidence retrieval, where we design a retrieval pipeline to collect and re-rank relevant documents from verifiable sources. Then, RAFTS forms contrastive arguments (i.e., supporting or refuting) conditioned on the retrieved evidence. In addition, RAFTS leverages an embedding model to identify informative demonstrations, followed by in-context prompting to generate the prediction and explanation. Our method effectively retrieves relevant documents as evidence and evaluates arguments from varying perspectives, incorporating nuanced information for fine-grained decision-making. Combined with informative in-context examples as prior, RAFTS achieves significant improvements to supervised and LLM baselines without complex prompts. We demonstrate the effectiveness of our method through extensive experiments, where RAFTS can outperform GPT-based methods with a significantly smaller 7B LLM.

cs.CL↗

arXiv2024

Two-Stage Adaptive Network for Semi-Supervised Cross-Domain Crater Detection under Varying Scenario Distributions

Crater detection can provide valuable information for humans to explore the topography and understand the history of extraterrestrial planets. Due to the significantly varying scenario distributions, existing detection models trained on known labelled crater datasets are hardly effective when applied to new unlabelled planets. To address this issue, we propose a two-stage adaptive network (TAN) for semi-supervised cross-domain crater detection. Our network is built on the YOLOv5 detector, where a series of strategies are employed to enhance its cross-domain generalisation ability. In the first stage, we propose an attention-based scale-adaptive fusion (ASAF) strategy to handle objects with significant scale variances. Furthermore, we propose a smoothing hard example mining (SHEM) loss function to address the issue of overfitting on hard examples. In the second stage, we propose a sort-based pseudo-labelling fine-tuning (SPF) strategy for semi-supervised learning to mitigate the distributional differences between source and target domains. For both stages, we employ weak or strong image augmentation to suit different cross-domain tasks. Experimental results on benchmark datasets demonstrate that the proposed network can enhance domain adaptation ability for crater detection under varying scenario distributions.

cs.CV↗

arXiv2024

LLMs for User Interest Exploration in Large-scale Recommendation Systems

Traditional recommendation systems are subject to a strong feedback loop by learning from and reinforcing past user-item interactions, which in turn limits the discovery of novel user interests. To address this, we introduce a hybrid hierarchical framework combining Large Language Models (LLMs) and classic recommendation models for user interest exploration. The framework controls the interfacing between the LLMs and the classic recommendation models through "interest clusters", the granularity of which can be explicitly determined by algorithm designers. It recommends the next novel interests by first representing "interest clusters" using language, and employs a fine-tuned LLM to generate novel interest descriptions that are strictly within these predefined clusters. At the low level, it grounds these generated interests to an item-level policy by restricting classic recommendation models, in this case a transformer-based sequence recommender to return items that fall within the novel clusters generated at the high level. We showcase the efficacy of this approach on an industrial-scale commercial platform serving billions of users. Live experiments show a significant increase in both exploration of novel interests and overall user enjoyment of the platform.

cs.IR↗

arXiv2024

SOMTP: Self-Supervised Learning-Based Optimizer for MPC-Based Safe Trajectory Planning Problems in Robotics

Model Predictive Control (MPC)-based trajectory planning has been widely used in robotics, and incorporating Control Barrier Function (CBF) constraints into MPC can greatly improve its obstacle avoidance efficiency. Unfortunately, traditional optimizers are resource-consuming and slow to solve such non-convex constrained optimization problems (COPs) while learning-based methods struggle to satisfy the non-convex constraints. In this paper, we propose SOMTP algorithm, a self-supervised learning-based optimizer for CBF-MPC trajectory planning. Specifically, first, SOMTP employs problem transcription to satisfy most of the constraints. Then the differentiable SLPG correction is proposed to move the solution closer to the safe set and is then converted as the guide policy in the following training process. After that, inspired by the Augmented Lagrangian Method (ALM), our training algorithm integrated with guide policy constraints is proposed to enable the optimizer network to converge to a feasible solution. Finally, experiments show that the proposed algorithm has better feasibility than other learning-based methods and can provide solutions much faster than traditional optimizers with similar optimality.

cs.RO↗

arXiv2024

ICGNet: A Unified Approach for Instance-Centric Grasping

Accurate grasping is the key to several robotic tasks including assembly and household robotics. Executing a successful grasp in a cluttered environment requires multiple levels of scene understanding: First, the robot needs to analyze the geometric properties of individual objects to find feasible grasps. These grasps need to be compliant with the local object geometry. Second, for each proposed grasp, the robot needs to reason about the interactions with other objects in the scene. Finally, the robot must compute a collision-free grasp trajectory while taking into account the geometry of the target object. Most grasp detection algorithms directly predict grasp poses in a monolithic fashion, which does not capture the composability of the environment. In this paper, we introduce an end-to-end architecture for object-centric grasping. The method uses pointcloud data from a single arbitrary viewing direction as an input and generates an instance-centric representation for each partially observed object in the scene. This representation is further used for object reconstruction and grasp detection in cluttered table-top scenes. We show the effectiveness of the proposed method by extensively evaluating it against state-of-the-art methods on synthetic datasets, indicating superior performance for grasping and reconstruction. Additionally, we demonstrate real-world applicability by decluttering scenes with varying numbers of objects.

cs.RO↗

arXiv2024

Photonic Millimeter-wave Generation Beyond the Cavity Thermal Limit

Next-generation communications, radar and navigation systems will extend and exploit the higher bandwidth of the millimeter-wave domain for increased communication data rates as well as radar with higher sensitivity and increased spatial resolution. However, realizing these advantages will require the generation of millimeter-wave signals with low phase noise in simple and compact form-factors. The rapidly developing field of photonic integration addresses this challenge and provides a path toward simplified and portable, low-noise mm-wave generation for these applications. We leverage these advances by heterodyning two silicon photonic chip lasers, phase-locked to the same miniature Fabry-Perot (F-P) cavity to demonstrate a simple framework for generating low-noise millimeter-waves with phase noise below the thermal limit of the F-P cavity. Specifically, we generate 94.5 GHz and 118.1 GHz millimeter-wave signals with phase noise of -117 dBc/Hz at 10 kHz offset, decreasing to -120 dBc/Hz at 40 kHz offset, a record low value for such photonic devices. We achieve this with existing technologies that can be integrated into a platform less than $\approx$ 10 mL in volume. Our work illustrates the significant potential and advantages of low size, weight, and power (SWaP) photonic-sourced mm-waves for communications and sensing.

physics.optics↗

arXiv2024

Scaling up Multi-domain Semantic Segmentation with Sentence Embeddings

We propose an approach to semantic segmentation that achieves state-of-the-art supervised performance when applied in a zero-shot setting. It thus achieves results equivalent to those of the supervised methods, on each of the major semantic segmentation datasets, without training on those datasets. This is achieved by replacing each class label with a vector-valued embedding of a short paragraph that describes the class. The generality and simplicity of this approach enables merging multiple datasets from different domains, each with varying class labels and semantics. The resulting merged semantic segmentation dataset of over 2 Million images enables training a model that achieves performance equal to that of state-of-the-art supervised methods on 7 benchmark datasets, despite not using any images therefrom. By fine-tuning the model on standard semantic segmentation datasets, we also achieve a significant improvement over the state-of-the-art supervised segmentation on NYUD-V2 and PASCAL-context at 60% and 65% mIoU, respectively. Based on the closeness of language embeddings, our method can even segment unseen labels. Extensive experiments demonstrate strong generalization to unseen image domains and unseen labels, and that the method enables impressive performance improvements in downstream applications, including depth estimation and instance segmentation.

cs.CV↗

arXiv2024

GPT4Motion: Scripting Physical Motions in Text-to-Video Generation via Blender-Oriented GPT Planning

Recent advances in text-to-video generation have harnessed the power of diffusion models to create visually compelling content conditioned on text prompts. However, they usually encounter high computational costs and often struggle to produce videos with coherent physical motions. To tackle these issues, we propose GPT4Motion, a training-free framework that leverages the planning capability of large language models such as GPT, the physical simulation strength of Blender, and the excellent image generation ability of text-to-image diffusion models to enhance the quality of video synthesis. Specifically, GPT4Motion employs GPT-4 to generate a Blender script based on a user textual prompt, which commands Blender's built-in physics engine to craft fundamental scene components that encapsulate coherent physical motions across frames. Then these components are inputted into Stable Diffusion to generate a video aligned with the textual prompt. Experimental results on three basic physical motion scenarios, including rigid object drop and collision, cloth draping and swinging, and liquid flow, demonstrate that GPT4Motion can generate high-quality videos efficiently in maintaining motion coherency and entity consistency. GPT4Motion offers new insights in text-to-video research, enhancing its quality and broadening its horizon for further explorations.

cs.CV↗

arXiv2024

Endora: Video Generation Models as Endoscopy Simulators

Generative models hold promise for revolutionizing medical education, robot-assisted surgery, and data augmentation for machine learning. Despite progress in generating 2D medical images, the complex domain of clinical video generation has largely remained untapped.This paper introduces \model, an innovative approach to generate medical videos that simulate clinical endoscopy scenes. We present a novel generative model design that integrates a meticulously crafted spatial-temporal video transformer with advanced 2D vision foundation model priors, explicitly modeling spatial-temporal dynamics during video generation. We also pioneer the first public benchmark for endoscopy simulation with video generation models, adapting existing state-of-the-art methods for this endeavor.Endora demonstrates exceptional visual quality in generating endoscopy videos, surpassing state-of-the-art methods in extensive testing. Moreover, we explore how this endoscopy simulator can empower downstream video analysis tasks and even generate 3D medical scenes with multi-view consistency. In a nutshell, Endora marks a notable breakthrough in the deployment of generative AI for clinical endoscopy research, setting a substantial stage for further advances in medical content generation. For more details, please visit our project page: https://endora-medvidgen.github.io/.

cs.CV↗

arXiv2024

CodeApex: A Bilingual Programming Evaluation Benchmark for Large Language Models

With the emergence of Large Language Models (LLMs), there has been a significant improvement in the programming capabilities of models, attracting growing attention from researchers. Evaluating the programming capabilities of LLMs is crucial as it reflects the multifaceted abilities of LLMs, and it has numerous downstream applications. In this paper, we propose CodeApex, a bilingual benchmark dataset focusing on the programming comprehension, code generation, and code correction abilities of LLMs. Programming comprehension task tests LLMs on multiple-choice exam questions covering conceptual understanding, commonsense reasoning, and multi-hop reasoning. The code generation task evaluates LLMs through completing C++ functions based on provided descriptions and prototypes. The code correction task asks LLMs to fix real-world erroneous code segments with different error messages. We evaluate 12 widely used LLMs, including both general-purpose and specialized models. GPT-4 exhibits the best programming capabilities, achieving approximate accuracy of 69%, 54%, and 66% on the three tasks, respectively. Compared to human performance, there is still significant room for improvement in LLM programming. We hope that CodeApex can serve as a reference for evaluating the coding capabilities of LLMs, further promoting their development and growth.

cs.CL↗

arXiv2024

Bond ordering in flux phases

Contrary to canonical expectations we show that lattice translational symmetry breaking often accompanies uniformly ordered flux phases. We demonstrate this phenomena by studying a spinless-fermion model on a square latttice with nearest-neighbor repulsion. We find an array of flux patterns, as a function interaction strength and filling factor, that break time-reversal symmetry but may or may not preserve translational symmetry. A key finding is that the pattern of flux ordering does not uniquely determine the electronic properties. As such a class of new phases of matter are introduced that expand the candidate ground states in many body interacting systems.

cond-mat.str-el↗