DR-NTU (Digital Repository of NTU)
Not a member yet
116018 research outputs found
Sort by
李贽人道思想的演变及其备受非难的原因——以《焚书》为中心 = The evolution of Li Zhi's humanistic thought and the reasons for its controversy: a study centered on A Book to Burn
No full text李贽是明末极具争议的思想家，其倡导真诚、强调个体自由、圣凡平等、反对伪道学等人道思想在思想史占据重要地位，却在当时被主流学界和统治者排挤。本文以《焚书》为研究对象，据时代背景、政治环境和个人经历探讨李贽人道思想的演变及备受责难的原因。
李贽的人道思想经历了从强调个体价值，到关注社会公平，再到激烈批判礼教和权威的转变。早期，他围绕个体修养和人格塑造展开，强调性高洁、务实的人伦原则、自然之性的发展、真诚等。晚年则日趋激进，提出圣凡平等、男女平等等观念，直接挑战明朝专制统治。
李贽的思想备受非难，除挑战传统秩序外，也因内部矛盾与建设性不足，缺乏完整的思想体系作为现实可行性的替代方案。研究李贽思想的演变，有助于理解明末思想转型的复杂性、重审“异端”思想的价值、思考社会秩序与思想解放间的平衡问题。
Li Zhi was a highly controversial late-Ming thinker whose advocacy of sincerity, individual freedom, equality between saints and commoners, and opposition to pseudo-Confucianism was influential in intellectual history but rejected by the mainstream and ruling authorities. This study examines A Book to Burn (Fen Shu) to explore the evolution of his humanistic thought and the reasons for its condemnation, considering historical, political, and personal contexts.
Li Zhi’s thought shifted from individual self-cultivation to concerns about social justice, eventually leading to a radical critique of Confucian orthodoxy and political authority. Early on, he emphasized moral integrity, pragmatic ethics, natural instincts, and sincerity. In his later years, he proposed equality between saints and commoners and gender equality, directly challenging Ming authoritarianism.
His ideas faced opposition not only for subverting tradition but also due to their internal contradictions and lack of a viable alternative system. While revolutionary, his thought lacked a systematic framework for practical implementation. Studying its evolution offers insight into late-Ming intellectual shifts and the role of “heretical” ideas in debates on social order and intellectual freedom.Bachelor's degre
Enhancing latent cross-attention-based diffusion models using graph neural networks for generating pocket-aware and target protein-specific novel therapeutic peptide designs
No full textWhile existing therapies ultimately face challenges, therapeutic peptides are designed to produce targeted biological effects, minimising off-target side effects and providing more tailored and cost-effective treatment options for a wide range of pathologies. However, designing them can be challenging due to a multitude of reasons, like the scarcity of peptide data. Several Deep Learning methods have been applied to reduce development costs and time. Nevertheless, these methods either require large datasets or struggle to incorporate the functional information necessary to generate therapeutic peptides that are structurally and functionally complementary to the target.
In our research project, we introduced a new method for generating novel, pocket-aware, target protein-specific peptide sequences by building upon the work of Sayuti et al. [36]. Our approach limits reliance on prior knowledge by integrating a latent diffusion model-centric architecture with an E(3)-invariant structure representation of the binding pocket extracted using a Graph Attention Network (GAT). The binding-site representation allows us to embed critical functional and structural information of the target protein’s pocket. Whereas, by introducing binding-site embeddings as constraints in the Cross-Attention block of the architecture, this approach allows us to indirectly capture multi-modal information and incorporate target specificity. The validity of our approach is demonstrated by its superior performance on evaluation metrics, outperforming those reported by Sayuti et al. [36], as well as models based on Variational Autoencoders (VAE) and Wasserstein Autoencoders (WAE). The results of the study underscore the importance of incorporating binding-site information in generative models to design peptides that are not only structurally and functionally relevant but also more likely to interact effectively with target proteins.Bachelor's degre
Optimized lyophilization protocol for better efficacy and storage capabilities of probiotic extracellular vesicles used in gastrointestinal disease treatment
No full textThe Nanoparticle drug delivery system is a current research area that many are working on to
explore the possible therapeutic capability for treatment. With the usage of nanoparticles, we are
hopeful of better drug delivery capabilities and a higher rate of treatment success. In this study, we
look at a specific type of nanoparticle, the cell-derived nanovesicles, and specifically for this
project, the Extracellular Vesicles (EVs) that are derived from the Lactobacillus rhamnosus
GG(LGG) probiotic. These would then be used in the treatment of Inflammatory Bowel disease
which comes in the form of either ulcerative colitis or Crohn’s disease.
However, the storage capability of such a novel treatment method is greatly compromised by the
delicate and intricate properties of the EVs which require it to be stored in specific conditions of
81 degrees Celsius. Therefore, lyophilization has been used to aid in this process. These EVs that
would be used in the drug formulation would then be treated with cryoprotectants, the
disaccharides of either sucrose or trehalose in the ratio of (1:100). These would then be put to a
storage test by leaving them in 2 different temperature conditions of 4℃, 25℃. Respectively to
study the effects of different storage environments and the viability of EVs. After which efficacy
tests were carried out to verify the effectiveness of these EVs being stored in different conditions.
This includes TEER and QPCR tests to determine whether the structural integrity is retained or
compromised, followed by detecting if there are immune response and immunomodulatory effects.
Results show that lyophilization with cryoprotectants was able to maintain the structural integrity
of the EVs while enhancing the storage capability of it. Sucrose specifically showed better results
in general compared to trehalose, which can translate to being the better lyophilization protocol to
be used to treat the probiotic EVs. For the aspect of therapeutic effects, sucrose-treated EVs
performed better in enduring a tighter barrier, while trehalose showed better results in eliciting an
immune response. The reason for such may be due to sugar toxicity which can be further studiedBachelor's degre
AI based chatbot
No full textAssistAI is an AI-powered chatbot designed to address the academic and research needs of
faculty members in higher education. Leveraging state-of-the-art Large Language Models
(LLMs), AssistAI simplifies complex academic workflows by automating research assistance,
lecture note generation, and assessment creation. The chatbot provides tailored solutions such as
automated paper discovery, content summaries, structured presentation outlines, and quiz
development with customizable difficulty levels. By integrating features like conversational
interfaces, proactive task management, and adaptive functionality, AssistAI enhances
productivity while minimizing the administrative burden on educators.Bachelor's degre
Formation control of autonomous vehicle platoons
No full textMulti-agent systems consist of multiple intelligent agents equipped with perception, computation, execution, and communication capabilities. Consequently, multi-agent cooperative control has garnered significant research interest. Formation control, a crucial branch of multi-agent cooperative control, aims to maintain or adjust formations of multiple agents according to task requirements. With the advancement of autonomous driving technology, vehicle platoon control is increasingly recognized for its potential in intelligent transportation systems.
This dissertation investigates the formation control problem of autonomous vehicle platoons, focusing on their mathematical modeling and distributed control methods. First, a mathematical model of the vehicle platoon is established under practical constraints. Based on this model, a distributed formation control algorithm is designed to ensure stable platooning under various initial conditions. Subsequently, numerical simulations are conducted using the MATLAB platform to evaluate the proposed algorithm’s convergence and stability. The simulation results demonstrate that the proposed algorithm effectively achieves stable formation control for autonomous vehicle platoons, verifying its feasibility.
Finally, this dissertation summarizes the findings and explores future research directions for consensus of formation control in multi-agent systems. The research contributes to the theoretical foundation and technical development of formation control in autonomous vehicle platoons.Master's degre
Exploring the utilization of tensor networks for quantum circuit simulation and its use case application
No full textQuantum Computing is an upcoming paradigm of computing that harnesses the laws
of quantum mechanics like superposition and entanglement to offer computational
speedup for certain challenging problems that current supercomputers cannot solve
efficiently and effectively. Challenges to achieving fault tolerant quantum computers
that makes this advantage practical have pushed scientists to explore the possibility of
simulating quantum computation on classical computers to determine how quantum
algorithms would behave on real quantum hardware devices. In this project, the goal
is exploration into how to simulate quantum computation using a special technique of
tensor networks which allows scalability in problem solving. Aiming at bridging the
gap between theory and application, the Variational Quantum Eigensolver algorithm is
implemented in Qibotn, a module for tensor network quantum simulation developed by
Institute of High Performance Computing, A*STAR. Further, a real-world, small-scale
application in logistics, namely the Vehicle Routing Problem is solved using quantum
simulation by encoding the problem statement into a Quadratic Unconstrained Binary
OptimizationwhichismodeledintoaquantumHamiltonian,whosegroundstateenergy
is calculated for by the algorithm, solves the given problem statement. Thus, in this
project we have implemented the VQE algorithm as a feature in Qibotn for Qibo users
and have built a working application based on tensor network quantum simulations to
demonstrate the usefulness of the algorithm built, lending it academic and practical
relevance in solving real-world problem using an emerging technology, which can be
scaled up in the future to solve for more complex scenarios using the computational
advantage offered by tensor networks in overcoming the challenge of simulating large
quantum circuits on classical machines.Bachelor's degre
Isotropic metamaterial stiffness beyond Hashin-Shtrikman upper bound
No full textSince its introduction more than 60 years ago, the Hashin-Shtrikman upper bound has stood as
the theoretical limit for the stiffness of isotropic composites and porous solids, acting as an
important reference against which the moduli of heterogeneous structural materials are
assessed. Here, we show through first-principles calculations, supported by finite element
simulations, that the Hashin-Shtrikman upper bound can be exceeded by the isotropic elastic
response of an anisotropic structure constructed from an anisotropic material. The material and
structural anisotropies mutually reinforce each other to realize the overall isotropic response,
without incurring the mass penalty faced by the hybridization of geometries with
complementary anisotropies. 3 designs were investigated (plate BCC, plate FCC and plate SC)
but only plate SC yielded a solution for the anisotropic properties of the material, which are
remarkably similar to that of single crystal nickel and single crystal ferrite.Economic Development Board (EDB)Nanyang Technological UniversitySubmitted/Accepted versionThis work was partially supported by C.Q.L’s Nanyang Assistant Professorship grant (award no.: #022081-00001) and EDB-OSTIn grant (award no.: S22-19018-STDP)
Optimized synthesis of metal-organic framework (MOF) based electro-catalyst for nitrogen reduction reaction (NRR)
No full textAmmonia (NH3), the critical compound found in fertilisers and many other industrial processes, has been environmentally taxing to reproduce and consumes large amounts of energy under the conventional Haber-Bosch process. This study investigates an alternative approach based on the method of electro-catalyst nitrogen reduction reaction (E-NRR) with Metal-Organic Frameworks (MOF) as the electrocatalyst. MIL-101 (Cr) is the chosen MOF for this study, owing to its good catalytic properties of high porosity, surface area, robust stability and tunability. The hydrothermal synthesis method was utilised in this study to synthesise chromium nitrate nonahydrate and terephthalic acid, forming MIL-101 (Cr). Purification, desiccation and characterisation methods are discussed within the experimental segment of this report. Characterisation results were studied to understand the structure and electronic properties of the MOF. The findings will be consolidated and discussed, elucidating the potential of MOF-based electrocatalyst, furthering the research of a decentralised ammonia production, laying additional groundwork for optimisation and scalability of MOF.Bachelor's degre
Spatio-temporal analysis of public transportation systems using network autoregressive models
No full textThis research study covers the literature behind statistical spatio-temporal analysis as well as the
progress made thus far in exploring its applications in the domain of public transportation. The
research project delves into data provided by Open Data NY covering the New York City (NYC)
Subway System, making use of statistical methods to accurately forecast future changes.
This analysis showcases the usage and application of the Generalised Network Autoregressive
(GNAR) model as proposed by authors Knight and Leeming, to fit a model to predict the ridership of
the various subway stations around NYC. Authors Nason and Wei showcased the model’s prowess in
the domain of economics by quantifying and predicting the economic response to COVID-19 across
various European countries. This paper showcases the GNAR model’s use case in the domain of
transport, making use of the subway station’s pre-existing graph network structure to aid in
predictions.
As it concludes, the paper would have gone through the steps taken to extract, transform, and load the
data making use of Python and the Pandas library. The raw data obtained using the MTA API will be
transformed into graphical networks and vector time series’ compatible with the autoregressive model.
Subsequent analyses would be conducted in R, outlining the conditions and consolations given to
provide a model that would be useful in forecasting. This includes an investigation to the Bayesian
Information Criterion (BIC) of each model, observations of its predictive power represented by the
mean squared prediction error (MSPE) and explained variance (R-squared) of each model.
The paper also covers several challenges faced during the data exploration of model fitting, offering
potential solutions to future renditions of this project. Specifically, improvements to parameter
selection and graph network construction have been outlined in the closing review, along with
extensions to compare regression-based analyses against machine learning and large language model
(LLM) based analyses.Bachelor's degre
Target driven visual navigation for a mobile robot using deep reinforcement learning
No full textTarget-driven visual navigation remains a critical challenge for autonomous mobile robots (AMRs) operating in dynamic, unstructured environments. Traditional approaches relying on pre-built maps or GPS-based localization often fail in GPS-denied indoor spaces or scenarios requiring adaptation to unseen layouts. This dissertation presents a novel deep reinforcement learning (DRL) framework that enables AMRs to navigate toward alphanumeric targets using egocentric visual inputs, eliminating dependency on prior environmental knowledge.
The proposed framework integrates three key innovations: (1) zero-shot object detection for robust localization of numeric targets without class-specific training, (2) Transformer-based Optical Character Recognition (TrOCR) for discriminative feature extraction, and (3) Principal Component Analysis (PCA) to enhance numeric differentiation by reducing redundant visual information. Leveraging procedural environment generation via ProcTHOR, we create diverse corridor configurations with varying lighting, textures, and obstacle layouts to ensure generalization. The navigation policy is optimized through Proximal Policy Optimization (PPO), combining sparse rewards for target proximity with penalties for inefficient movements.
Experimental evaluations demonstrate that the proposed model achieves a 70% success rate in target navigation tasks, marking a significant breakthrough compared to baseline models that lack alphanumeric image recognition capabilities. This performance gap highlights the critical role of integrating visual-textual understanding for navigating alphanumeric targets.Master's degre