arXiv subjects

Shikun Liu

Publications and source records attributed to Shikun Liu.

31 records · Page 2Linked to original sources

arXiv2024

Prismer: A Vision-Language Model with Multi-Task Experts

Recent vision-language models have shown impressive multi-modal generation capabilities. However, typically they require training huge models on massive datasets. As a more scalable alternative, we introduce Prismer, a data- and parameter-efficient vision-language model that leverages an ensemble of task-specific experts. Prismer only requires training of a small number of components, with the majority of network weights inherited from multiple readily-available, pre-trained experts, and kept frozen during training. By leveraging experts from a wide range of domains, we show Prismer can efficiently pool this expert knowledge and adapt it to various vision-language reasoning tasks. In our experiments, we show that Prismer achieves fine-tuned and few-shot learning performance which is competitive with current state-of-the-arts, whilst requiring up to two orders of magnitude less training data. Code is available at https://github.com/NVlabs/prismer.

cs.LG↗

arXiv2023

Nonlinear interaction of head$-$on solitary waves in integrable and nonintegrable systems

This study numerically investigates the nonlinear interaction of head-on solitary waves in a granular chain (a nonintegrable system) and compares the simulation results with the theoretical results in fluid (an integrable system). Three stages (i.e., pre-in-phase traveling stage, central-collision stage, and post-in-phase traveling stage) are identified to describe the nonlinear interaction processes in the granular chain. The nonlinear scattering effect occurs in the central-collision stage, which decreases the amplitude of incident solitary waves. Compared with the leading-time phase in the incident and separation collision processes, the lagging-time phase in the separation collision process is smaller. This asymmetrical nonlinear collision results in an occurrence of leading phase shifts of time and space in the post-in-phase traveling stage. We next find that solitary wave amplitude does not influence the immediate space-phase shift in the granular chain. The space$-$phase shift of the post-in-phase traveling stage is only determined by measurement position rather than wave amplitude. The results are reversed in the fluid. An increase in solitary wave amplitude leads to decreased attachment, detachment and residence times for granular chain and fluid. For the immediate time-phase shift, leading and lagging phenomena appear in the granular chain and the fluid, respectively. These results offer new knowledge for designing mechanical metamaterials and energy-mitigating systems.

cond-mat.soft↗

arXiv2023

Structural Re-weighting Improves Graph Domain Adaptation

In many real-world applications, graph-structured data used for training and testing have differences in distribution, such as in high energy physics (HEP) where simulation data used for training may not match real experiments. Graph domain adaptation (GDA) is a method used to address these differences. However, current GDA primarily works by aligning the distributions of node representations output by a single graph neural network encoder shared across the training and testing domains, which may often yield sub-optimal solutions. This work examines different impacts of distribution shifts caused by either graph structure or node attributes and identifies a new type of shift, named conditional structure shift (CSS), which current GDA approaches are provably sub-optimal to deal with. A novel approach, called structural reweighting (StruRW), is proposed to address this issue and is tested on synthetic graphs, four benchmark datasets, and a new application in HEP. StruRW has shown significant performance improvement over the baselines in the settings with large graph structure shifts, and reasonable performance improvement when node attribute shift dominates.

cs.LG↗

arXiv2023

vMAP: Vectorised Object Mapping for Neural Field SLAM

We present vMAP, an object-level dense SLAM system using neural field representations. Each object is represented by a small MLP, enabling efficient, watertight object modelling without the need for 3D priors. As an RGB-D camera browses a scene with no prior information, vMAP detects object instances on-the-fly, and dynamically adds them to its map. Specifically, thanks to the power of vectorised training, vMAP can optimise as many as 50 individual objects in a single scene, with an extremely efficient training speed of 5Hz map update. We experimentally demonstrate significantly improved scene-level and object-level reconstruction quality compared to prior neural field SLAM systems. Project page: https://kxhit.github.io/vMAP.

cs.CV↗

arXiv2022

Auto-Lambda: Disentangling Dynamic Task Relationships

Understanding the structure of multiple related tasks allows for multi-task learning to improve the generalisation ability of one or all of them. However, it usually requires training each pairwise combination of tasks together in order to capture task relationships, at an extremely high computational cost. In this work, we learn task relationships via an automated weighting framework, named Auto-Lambda. Unlike previous methods where task relationships are assumed to be fixed, Auto-Lambda is a gradient-based meta learning framework which explores continuous, dynamic task relationships via task-specific weightings, and can optimise any choice of combination of tasks through the formulation of a meta-loss; where the validation loss automatically influences task weightings throughout training. We apply the proposed framework to both multi-task and auxiliary learning problems in computer vision and robotics, and show that Auto-Lambda achieves state-of-the-art performance, even when compared to optimisation strategies designed specifically for each problem and data domain. Finally, we observe that Auto-Lambda can discover interesting learning behaviors, leading to new insights in multi-task learning. Code is available at https://github.com/lorenmt/auto-lambda.

cs.LG↗

arXiv2022

Semi-supervised Graph Neural Networks for Pileup Noise Removal

The high instantaneous luminosity of the CERN Large Hadron Collider leads to multiple proton-proton interactions in the same or nearby bunch crossings (pileup). Advanced pileup mitigation algorithms are designed to remove this noise from pileup particles and improve the performance of crucial physics observables. This study implements a semi-supervised graph neural network for particle-level pileup noise removal, by identifying individual particles produced from pileup. The graph neural network is firstly trained on charged particles with known labels, which can be obtained from detector measurements on data or simulation, and then inferred on neutral particles for which such labels are missing. This semi-supervised approach does not depend on the ground truth information from simulation and thus allows us to perform training directly on experimental data. The performance of this approach is found to be consistently better than widely-used domain algorithms and comparable to the fully-supervised training using simulation truth information. The study serves as the first attempt at applying semi-supervised learning techniques to pileup mitigation, and opens up a new direction of fully data-driven machine learning pileup mitigation studies.

hep-ex↗

arXiv2022

Bootstrapping Semantic Segmentation with Regional Contrast

We present ReCo, a contrastive learning framework designed at a regional level to assist learning in semantic segmentation. ReCo performs semi-supervised or supervised pixel-level contrastive learning on a sparse set of hard negative pixels, with minimal additional memory footprint. ReCo is easy to implement, being built on top of off-the-shelf segmentation networks, and consistently improves performance in both semi-supervised and supervised semantic segmentation methods, achieving smoother segmentation boundaries and faster convergence. The strongest effect is in semi-supervised learning with very few labels. With ReCo, we achieve high-quality semantic segmentation models, requiring only 5 examples of each semantic class. Code is available at https://github.com/lorenmt/reco.

cs.CV↗

arXiv2021

iMAP: Implicit Mapping and Positioning in Real-Time

We show for the first time that a multilayer perceptron (MLP) can serve as the only scene representation in a real-time SLAM system for a handheld RGB-D camera. Our network is trained in live operation without prior data, building a dense, scene-specific implicit 3D model of occupancy and colour which is also immediately used for tracking. Achieving real-time SLAM via continual training of a neural network against a live image stream requires significant innovation. Our iMAP algorithm uses a keyframe structure and multi-processing computation flow, with dynamic information-guided pixel sampling for speed, with tracking at 10 Hz and global map updating at 2 Hz. The advantages of an implicit MLP over standard dense SLAM techniques include efficient geometry representation with automatic detail control and smooth, plausible filling-in of unobserved regions such as the back surfaces of objects.

cs.CV↗

arXiv2020

Shape Adaptor: A Learnable Resizing Module

We present a novel resizing module for neural networks: shape adaptor, a drop-in enhancement built on top of traditional resizing layers, such as pooling, bilinear sampling, and strided convolution. Whilst traditional resizing layers have fixed and deterministic reshaping factors, our module allows for a learnable reshaping factor. Our implementation enables shape adaptors to be trained end-to-end without any additional supervision, through which network architectures can be optimised for each individual task, in a fully automated way. We performed experiments across seven image classification datasets, and results show that by simply using a set of our shape adaptors instead of the original resizing layers, performance increases consistently over human-designed networks, across all datasets. Additionally, we show the effectiveness of shape adaptors on two other applications: network compression and transfer learning. The source code is available at: https://github.com/lorenmt/shape-adaptor.

cs.LG↗

arXiv2019

Self-Supervised Generalisation with Meta Auxiliary Learning

Learning with auxiliary tasks can improve the ability of a primary task to generalise. However, this comes at the cost of manually labelling auxiliary data. We propose a new method which automatically learns appropriate labels for an auxiliary task, such that any supervised learning task can be improved without requiring access to any further data. The approach is to train two neural networks: a label-generation network to predict the auxiliary labels, and a multi-task network to train the primary task alongside the auxiliary task. The loss for the label-generation network incorporates the loss of the multi-task network, and so this interaction between the two networks can be seen as a form of meta learning with a double gradient. We show that our proposed method, Meta AuXiliary Learning (MAXL), outperforms single-task learning on 7 image datasets, without requiring any additional data. We also show that MAXL outperforms several other baselines for generating auxiliary labels, and is even competitive when compared with human-defined auxiliary labels. The self-supervised nature of our method leads to a promising new direction towards automated generalisation. Source code can be found at https://github.com/lorenmt/maxl.

cs.LG↗

arXiv2019

End-to-End Multi-Task Learning with Attention

We propose a novel multi-task learning architecture, which allows learning of task-specific feature-level attention. Our design, the Multi-Task Attention Network (MTAN), consists of a single shared network containing a global feature pool, together with a soft-attention module for each task. These modules allow for learning of task-specific features from the global features, whilst simultaneously allowing for features to be shared across different tasks. The architecture can be trained end-to-end and can be built upon any feed-forward neural network, is simple to implement, and is parameter efficient. We evaluate our approach on a variety of datasets, across both image-to-image predictions and image classification tasks. We show that our architecture is state-of-the-art in multi-task learning compared to existing methods, and is also less sensitive to various weighting schemes in the multi-task loss function. Code is available at https://github.com/lorenmt/mtan.

cs.CV↗

arXiv2018

Learning a Hierarchical Latent-Variable Model of 3D Shapes

We propose the Variational Shape Learner (VSL), a generative model that learns the underlying structure of voxelized 3D shapes in an unsupervised fashion. Through the use of skip-connections, our model can successfully learn and infer a latent, hierarchical representation of objects. Furthermore, realistic 3D objects can be easily generated by sampling the VSL's latent probabilistic manifold. We show that our generative model can be trained end-to-end from 2D images to perform single image 3D model retrieval. Experiments show, both quantitatively and qualitatively, the improved generalization of our proposed model over a range of tasks, performing better or comparable to various state-of-the-art alternatives.

cs.CV↗

arXiv2016

A Note on Weighted Fermat Problem in Lp Space

In this paper, we begin by introducing a well-known geometry concept: the Fermat point in a triangle. Then, we generalize the problem and propose an iterative algorithm based on gradient descent to the weighted form in Lp space. We also build specific solutions to some special norms: one, two and infinity. We show that the solution may not be unique in norm-one and infinity. Finally, we provide our qualitative results in the PASCAL-based program.

math.OC↗