Computer Sciences Commons^™

A Black-Box Attack On Code Models Via Representation Nearest Neighbor Search, Jie Zhang, Wei Ma, Qiang Hu, Shangqing Liu, Xiaofei Xie, Yves Le Traon, Yang Liu

Research Collection School Of Computing and Information Systems

Existing methods for generating adversarial code examples face several challenges: limted availability of substitute variables, high verification costs for these substitutes, and the creation of adversarial samples with noticeable perturbations. To address these concerns, our proposed approach, RNNS, uses a search seed based on historical attacks to find potential adversarial substitutes. Rather than directly using the discrete substitutes, they are mapped to a continuous vector space using a pre-trained variable name encoder. Based on the vector representation, RNNS predicts and selects better substitutes for attacks. We evaluated the performance of RNNS across six coding tasks encompassing three programming languages: Java, …

Llm-Adapters: An Adapter Family For Parameter-Efficient Fine-Tuning Of Large Language Models, Zhiqiang Hu, Lei Wang, Yihuai Lan, Wanyu Xu, Ee-Peng Lim, Lidong Bing, Xing Xu, Soujanya Poria, Roy Ka-Wei Lee

Research Collection School Of Computing and Information Systems

The success of large language models (LLMs), like GPT-4 and ChatGPT, has led to the development of numerous cost-effective and accessible alternatives that are created by finetuning open-access LLMs with task-specific data (e.g., ChatDoctor) or instruction data (e.g., Alpaca). Among the various fine-tuning methods, adapter-based parameter-efficient fine-tuning (PEFT) is undoubtedly one of the most attractive topics, as it only requires fine-tuning a few external parameters instead of the entire LLMs while achieving comparable or even better performance. To enable further research on PEFT methods of LLMs, this paper presents LLMAdapters, an easy-to-use framework that integrates various adapters into LLMs and …

Examining The Inter-Consistency Of Large Language Models: An In-Depth Analysis Via Debate, Kai Xiong, Xiao Ding, Yixin Cao, Ting Liu, Bing Qin

Research Collection School Of Computing and Information Systems

Large Language Models (LLMs) have shown impressive capabilities in various applications, but they still face various inconsistency issues. Existing works primarily focus on the inconsistency issues within a single LLM, while we complementarily explore the inter-consistency among multiple LLMs for collaboration. To examine whether LLMs can collaborate effectively to achieve a consensus for a shared goal, we focus on commonsense reasoning, and introduce a formal debate framework (FORD) to conduct a three-stage debate among LLMs with real-world scenarios alignment: fair debate, mismatched debate, and roundtable debate. Through extensive experiments on various datasets, LLMs can effectively collaborate to reach a consensus …

Kape: Knn-Based Performance Testing For Deep Code Search, Yuejun Guo, Qiang Hu, Xiaofei Xie, Cordy Maxime, Mike Papadakis, Yves Le Traon

Research Collection School Of Computing and Information Systems

Code search is a common yet important activity of software developers. An efficient code search model can largely facilitate the development process and improve the programming quality. Given the superb performance of learning the contextual representations, deep learning models, especially pre-trained language models, have been widely explored for the code search task. However, studies mainly focus on proposing new architectures for ever-better performance on designed test sets but ignore the performance on unseen test data where only natural language queries are available. The same problem in other domains, e.g., CV and NLP, is usually solved by test input selection that …

Attack Prompt Generation For Red Teaming And Defending Large Language Models, Boyi Deng, Wenjie Wang, Fuli Feng, Yang Deng, Qifan Wang, Xiangnan He

Research Collection School Of Computing and Information Systems

Large language models (LLMs) are susceptible to red teaming attacks, which can induce LLMs to generate harmful content. Previous research constructs attack prompts via manual or automatic methods, which have their own limitations on construction cost and quality. To address these issues, we propose an integrated approach that combines manual and automatic methods to economically generate high-quality attack prompts. Specifically, considering the impressive capabilities of newly emerged LLMs, we propose an attack framework to instruct LLMs to mimic human-generated prompts through in-context learning. Furthermore, we propose a defense framework that fine-tunes victim LLMs through iterative interactions with the attack framework …

Disentangling Transformer Language Models As Superposed Topic Models, Jia Peng Lim, Hady Wirawan Lauw

Research Collection School Of Computing and Information Systems

Topic Modelling is an established research area where the quality of a given topic is measured using coherence metrics. Often, we infer topics from Neural Topic Models (NTM) by interpreting their decoder weights, consisting of top-activated words projected from individual neurons. Transformer-based Language Models (TLM) similarly consist of decoder weights. However, due to its hypothesised superposition properties, the final logits originating from the residual path are considered uninterpretable. Therefore, we posit that we can interpret TLM as superposed NTM by proposing a novel weight-based, model-agnostic and corpus-agnostic approach to search and disentangle decoder-only TLM, potentially mapping individual neurons to multiple …

Large Language Models As Source Planner For Personalized Knowledge-Grounded Dialogues, Hongru Wang, Minda Hu, Yang Deng, Rui Wang, Fei Mi, Weichao Wang, Yasheng Wang, Wai-Chung Kwan, Irwin King, Kam-Fai Wong

Research Collection School Of Computing and Information Systems

Open-domain dialogue system usually requires different sources of knowledge to generate more informative and evidential responses. However, existing knowledge-grounded dialogue systems either focus on a single knowledge source or overlook the dependency between multiple sources of knowledge, which may result in generating inconsistent or even paradoxical responses. To incorporate multiple knowledge sources and dependencies between them, we propose SAFARI, a novel framework that leverages the exceptional capabilities of large language models (LLMs) in planning, understanding, and incorporating under both supervised and unsupervised settings. Specifically, SAFARI decouples the knowledge grounding into multiple sources and response generation, which allows easy extension to …

Large Language Model Is Not A Good Few-Shot Information Extractor, But A Good Reranker For Hard Samples!, Yubo Ma, Yixin Cao, Yongchin Hong, Aixin Sun

Research Collection School Of Computing and Information Systems

Large Language Models (LLMs) have made remarkable strides in various tasks. However, whether they are competitive few-shot solvers for information extraction (IE) tasks and surpass fine-tuned small Pre-trained Language Models (SLMs) remains an open problem. This paper aims to provide a thorough answer to this problem, and moreover, to explore an approach towards effective and economical IE systems that combine the strengths of LLMs and SLMs. Through extensive experiments on nine datasets across four IE tasks, we show that LLMs are not effective few-shot information extractors in general, given their unsatisfactory performance in most settings and the high latency and …

Benchmarking Foundation Models With Language-Model-As-An-Examiner, Yushi Bai, Jiahao Ying, Yixin Cao, Xin Lv, Yuze He, Xiaozhi Wang, Jifan Yu, Kaisheng Zeng, Yijia Xiao, Haozhe Lyu, Jiayin Zhang, Juanzi Li, Lei Hou

Research Collection School Of Computing and Information Systems

Numerous benchmarks have been established to assess the performance of foundation models on open-ended question answering, which serves as a comprehensive test of a model’s ability to understand and generate language in a manner similar to humans. Most of these works focus on proposing new datasets, however, we see two main issues within previous benchmarking pipelines, namely testing leakage and evaluation automation. In this paper, we propose a novel benchmarking framework, Language-Model-as-an-Examiner, where the LM serves as a knowledgeable examiner that formulates questions based on its knowledge and evaluates responses in a reference-free manner. Our framework allows for effortless extensibility …

Molca: Molecular Graph-Language Modeling With Cross-Modal Projector And Uni-Modal Adapter, Zhiyuan Liu, Sihang Li, Yanchen Luo, Hao Fei, Yixin Cao, Kenji Kawaguchi, Xiang Wang, Tat-Seng Chua

Research Collection School Of Computing and Information Systems

Language Models (LMs) have demonstrated impressive molecule understanding ability on various 1D text-related tasks. However, they inherently lack 2D graph perception — a critical ability of human professionals in comprehending molecules’ topological structures. To bridge this gap, we propose MolCA: Molecular Graph-Language Modeling with Cross-Modal Projector and Uni-Modal Adapter. MolCA enables an LM (i.e., Galactica) to understand both text- and graph-based molecular contents via the cross-modal projector. Specifically, the cross-modal projector is implemented as a QFormer to connect a graph encoder’s representation space and an LM’s text space. Further, MolCA employs a uni-modal adapter (i.e., LoRA) for the LM’s efficient …

Random Variable Spaces: Mathematical Properties And An Extension To Programming Computable Functions, Mohammed Kurd-Misto

Computational and Data Sciences (PhD) Dissertations

This dissertation aims to extend the boundaries of Programming Computable Functions (PCF) by introducing a novel collection of categories referred to as Random Variable Spaces. Originating as a generalization of Quasi-Borel Spaces, Random Variable Spaces are rigorously defined as categories where objects are sets paired with a collection of random variables from an underlying measurable space. These spaces offer a theoretical foundation for extending PCF to natively handle stochastic elements.

The dissertation is structured into seven chapters that provide a multi-disciplinary background, from PCF and Measure Theory to Category Theory with special attention to Monads and the Giry Monad. The …

Hypothyroid Disease Analysis By Using Machine Learning, Sanjana Seelam

Electronic Theses, Projects, and Dissertations

Thyroid illness frequently manifests as hypothyroidism. It is evident that people with hypothyroidism are primarily female. Because the majority of people are unaware of the illness, it is quickly becoming more serious. It is crucial to catch it early on so that medical professionals can treat it more effectively and prevent it from getting worse. Machine learning illness prediction is a challenging task. Disease prediction is aided greatly by machine learning. Once more, unique feature selection strategies have made the process of disease assumption and prediction easier. To properly monitor and cure this illness, accurate detection is essential. In order …

Effective And Efficient Semantic Representations And Their Applications, Chong Cher Chia

Dissertations and Theses Collection (Open Access)

The proliferation of affordable and compact digital storage has also led to the creation of enormous databases of information, and much attention has been focused on the problem of processing unorganized and unstructured information into some form from which additional value can be extracted. Contemporary approaches to this problem virtually necessitate the use of complex models running on computational systems due to the sheer volume of information to be processed. While it is possible for the model to be fed the actual data as input, typically a representation of the data is used instead. These representations are therefore of interest, …

Towards Llm-Based Fact Verification On News Claims With A Hierarchical Step-By-Step Prompting Method, Xuan Zhang, Wei Gao

Research Collection School Of Computing and Information Systems

While large pre-trained language models (LLMs) have shown their impressive capabilities in various NLP tasks, they are still underexplored in the misinformation domain. In this paper, we examine LLMs with in-context learning (ICL) for news claim verification, and find that only with 4-shot demonstration examples, the performance of several prompting methods can be comparable with previous supervised models. To further boost performance, we introduce a Hierarchical Step-by-Step (HiSS) prompting method which directs LLMs to separate a claim into several subclaims and then verify each of them via multiple questionsanswering steps progressively. Experiment results on two public misinformation datasets show that …

Cgt-Gan: Clip-Guided Text Gan For Image Captioning, Jiarui Yu, Haoran Li, Yanbin Hao, Bin Zhu, Tong Xu, Xiangnan He

Research Collection School Of Computing and Information Systems

The large-scale visual-language pre-trained model, Contrastive Language-Image Pre-training (CLIP), has significantly improved image captioning for scenarios without human-annotated image-caption pairs. Recent advanced CLIP-based image captioning without human annotations follows a text-only training paradigm, i.e., reconstructing text from shared embedding space. Nevertheless, these approaches are limited by the training/inference gap or huge storage requirements for text embeddings. Given that it is trivial to obtain images in the real world, we propose CLIP-guided text GAN (CgT-GAN), which incorporates images into the training process to enable the model to "see" real visual modality. Particularly, we use adversarial training to teach CgT-GAN to mimic …

Teacher Candidates’ Conceptions And Practices Of Computational Thinking For Equity, Heather F. Clark, Symone A. Gyles, Imelda Nava-Landeros

Journal of Computer Science Integration

This study documents novice science and math teachers’ developing pedagogical approaches to integrating computational thinking (CT) and data into their courses to support educational equity and social justice. The 10 novice teacher candidates (TCs) studied were part of an urban teacher residency program that empowered them with an asset-based pedagogy we describe as “CT for Equity.” Drawing on coursework and interviews as data, we asked three questions: What are teachers’ conceptions of CT? What are their CT instructional practices? And how did their students respond to those practices? To explore conceptions of CT, we used Kafai et al.’s (2020) articulation …

Hallucination Detection: Robustly Discerning Reliable Answers In Large Language Models, Yuyuan Chen, Qiang Fu, Yichen Yuan, Zhihao Wen, Ge Fan, Dayiheng Liu, Dongmei Zhang, Zhixu Li, Yanghua Xiao

Research Collection School Of Computing and Information Systems

Large language models (LLMs) have gained widespread adoption in various natural language processing tasks, including question answering and dialogue systems. However, a major drawback of LLMs is the issue of hallucination, where they generate unfaithful or inconsistent content that deviates from the input source, leading to severe consequences. In this paper, we propose a robust discriminator named RelD to effectively detect hallucination in LLMs' generated answers. RelD is trained on the constructed RelQA, a bilingual question-answering dialogue dataset along with answers generated by LLMs and a comprehensive set of metrics. Our experimental results demonstrate that the proposed RelD successfully detects …

Using The Typescript Compiler To Fix Erroneous Node.Js Snippets, Brittany Reid, Christoph Treude, Markus Wagner

Research Collection School Of Computing and Information Systems

Most online code snippets do not run. This means that developers looking to reuse code from online sources must manually find and fix errors. We present an approach for automatically evaluating and correcting errors in Node.js code snippets: Node Code Correction (NCC). NCC leverages the ability of the TypeScript compiler to generate errors and inform code corrections through the combination of TypeScript’s builtin codefixes, our own targeted fixes, and deletion of erroneous lines. Compared to existing approaches using linters, our findings suggest that NCC is capable of detecting a larger number of errors per snippet and more error types, and …

Evaluating A Large Language Model’S Ability To Solve Programming Exercises From An Introductory Bioinformatics Course, Stephen R. Piccolo, Paul Denny, Andrew Luxton-Reilly, Samuel H. Payne, Perry G. Ridge

Faculty Publications

Life scientists frequently write computer code when doing research. Computer programming can aid researchers in performing tasks that are not supported by existing tools. Programming can also help researchers to implement analytical logic in a way that documents their steps and thus enables others to repeat those steps. Many educational resources are available to teach computer programming, but this skill remains challenging for many researchers and students to master. Artificial-intelligence tools like OpenAI’s ChatGPT are able to interpret human-language requests to generate code. Accordingly, we evaluated the extent to which this technology might be used to perform programming tasks described …

Asset-Based Approaches To Multilingual Students’ Computer Science Identity Development, Sharin Rawhiya Jacob, Mark Warschauer

Journal of Computer Science Integration

While computer science identity development has been examined in several studies, there is much to learn about the development of multilingual students’ computer science (CS) identities. To develop strong CS identities, multilingual students must engage in culturally and linguistically sustaining curriculum, pedagogy, and interaction that draws from their rich and varied resources. This theoretical paper is grounded in a justice-centered, asset-based framework that views the traditions and practices in students’ cultures and communities as strong contributors to knowledge construction in STEM. We draw on multiple studies exploring multilingual student CS identity development to better understand how their personal, familial, community-based, …

Wind River Elementary Computer Science Collaborative: Connecting Computer Science And Indigenous Identities And Knowledges On The Wind River Reservation, Joseph P. Wilson, Kathryn M. Rich, Jared O'Leary, Veronica Miller

Journal of Computer Science Integration

Three Northern Arapaho and Eastern Shoshone–serving districts formed a researcher–practitioner partnership with the Wyoming Department of Education, the American Institutes for Research®, and BootUp Professional Development to advance the computer science (CS) education of their elementary students in ways that strengthen their Indigenous identities and knowledges. In this paper, we share experiences from 2019 to 2022 with our curriculum development, professional development (PD), and classroom implementation. The researcher–practitioner partnership developed student and teacher materials to support elementary CS lessons aligned to Wyoming’s CS standards and “Indian Education for All” social studies standards. Indigenous community members served as experts to codesign …

Near Peer Pedagogies For Culturally Sustaining Computer Science And Mathematics, Cliff Freeman, Eli Tucker-Raymond, Juan Gutierrez

Journal of Computer Science Integration

In this study, College STEM Literacy Workers act as co-teachers in 9th and 10th grade classrooms alongside mathematics teachers as they both learn to integrate a 30-hour integrated computer science/mathematics curriculum. As part of the curriculum for educator learning, we adapted the Algebra Project’s “Model of Excellence” for culturally relevant-sustaining pedagogies. We used the framework to explore how teachers understand College STEM Literacy Workers’ contributions in the model as well as College STEM Literacy Workers’ own experiences in the classroom. We found that while teachers and College STEM Literacy Workers did not individually instantiate the model of excellence, they were …

Towards Safe Automated Refactoring Of Imperative Deep Learning Programs To Graph Execution, Raffi Khatchadourian, Tatiana Castro Vélez, Mehdi Bagherzadeh, Nan Jia, Anita Raja

Publications and Research

Efficiency is essential to support responsiveness w.r.t. ever-growing datasets, especially for Deep Learning (DL) systems. DL frameworks have traditionally embraced deferred execution-style DL code—supporting symbolic, graph-based Deep Neural Network (DNN) computation. While scalable, such development is error-prone, non-intuitive, and difficult to debug. Consequently, more natural, imperative DL frameworks encouraging eager execution have emerged at the expense of run-time performance. Though hybrid approaches aim for the “best of both worlds,” using them effectively requires subtle considerations to make code amenable to safe, accurate, and efficient graph execution. We present our ongoing work on automated refactoring that assists developers in specifying whether …

Towards Safe Automated Refactoring Of Imperative Deep Learning Programs To Graph Execution, Raffi Khatchadourian, Tatiana Castro Vélez, Mehdi Bagherzadeh, Nan Jia, Anita Raja

Publications and Research

Efficiency is essential to support responsiveness w.r.t. ever-growing datasets, especially for Deep Learning (DL) systems. DL frameworks have traditionally embraced deferred execution-style DL code—supporting symbolic, graph-based Deep Neural Network (DNN) computation. While scalable, such development is error-prone, non-intuitive, and difficult to debug. Consequently, more natural, imperative DL frameworks encouraging eager execution have emerged at the expense of run-time performance. Though hybrid approaches aim for the "best of both worlds," using them effectively requires subtle considerations to make code amenable to safe, accurate, and efficient graph execution. We present our ongoing work on automated refactoring that assists developers in specifying whether …

K-St: A Formal Executable Semantics Of The Structured Text Language For Plcs, Kun Wang, Jingyi Wang, Christopher M. Poskitt, Xiangxiang Chen, Jun Sun, Peng Cheng

Research Collection School Of Computing and Information Systems

Programmable Logic Controllers (PLCs) are responsible for automating process control in many industrial systems (e.g. in manufacturing and public infrastructure), and thus it is critical to ensure that they operate correctly and safely. The majority of PLCs are programmed in languages such as Structured Text (ST). However, a lack of formal semantics makes it difficult to ascertain the correctness of their translators and compilers, which vary from vendor-to-vendor. In this work, we develop K-ST, a formal executable semantics for ST in the K framework. Defined with respect to the IEC 61131-3 standard and PLC vendor manuals, K-ST is a high-level …

Are We Ready To Embrace Generative Ai For Software Q&A?, Bowen Xu, Thanh-Dat Nguyen, Thanh Le Cong, Thong Hoang, Jiakun Liu, Kisub Kim, Chen Gong, Changan Niu, Chenyu Wang, Xuan-Bach Dinh Le, David Lo

Research Collection School Of Computing and Information Systems

Stack Overflow, the world's largest software Q&A (SQA) website, is facing a significant traffic drop due to the emergence of generative AI techniques. ChatGPT is banned by Stack Overflow after only 6 days from its release. The main reason provided by the official Stack Overflow is that the answers generated by ChatGPT are of low quality. To verify this, we conduct a comparative evaluation of human-written and ChatGPT-generated answers. Our methodology employs both automatic comparison and a manual study. Our results suggest that human-written and ChatGPT-generated answers are semantically similar, however, human-written answers outperform ChatGPT-generated ones consistently across multiple aspects, …

Application Of Machine Learning Algorithms For Elucidation Of Biological Networks From Time Series Gene Expression Data, Krupa Nagori

Computational and Data Sciences (PhD) Dissertations

This dissertation provides a deep dive into understanding gene expression, interaction, regulation, and the intricate mechanisms behind heliotropism and phototropism. Additionally, the research accentuates the significance of machine learning techniques, specifically for gene regulatory networks (GRNs).

Chapter 1 offers an exhaustive benchmarking of GRN methodologies, furthering our comprehension of machine-learning models relevant to GRNs. The evaluation revealed that GRNTE, SWING, and BiXGBoost emerged as top-performing methods in GRN inference. The suitability of these models varies depending on specific research criteria such as computational needs, dataset dimensions, and performance metric emphasis. An innovation of this chapter was the introduction of Colab …

Restaurant Management Website, Akhil Sai Gollapudi

Electronic Theses, Projects, and Dissertations

In the ever-evolving corporate landscape of today, it is crucial to respond to customer needs as efficiently and in a timely manner as possible. The project's primary objective is to create a method for clients to make reservations for restaurants online. This makes life easier for busy customers in their daily life.

Today, people are looking for comfort thanks to rapidly developing technology. As we can see, people invent and implement new technologies in all fields according to customer needs.

We got a unique restaurant idea that helps people save time. People prefer quick methods to get things done. With …

Diffusion Variational Autoencoder For Tackling Stochasticity In Multi-Step Regression Stock Price Prediction, Kelvin J.L. Koa, Yunshan Ma, Ritchie Ng, Tat‑Seng Chua

Research Collection School Of Computing and Information Systems

Multi-step stock price prediction over a long-term horizon is crucial for forecasting its volatility, allowing financial institutions to price and hedge derivatives, and banks to quantify the risk in their trading books. Additionally, most financial regulators also require a liquidity horizon of several days for institutional investors to exit their risky assets, in order to not materially affect market prices. However, the task of multi-step stock price prediction is challenging, given the highly stochastic nature of stock data. Current solutions to tackle this problem are mostly designed for single-step, classification-based predictions, and are limited to low representation expressiveness. The problem …

Decompiling X86 Deep Neural Network Executables, Zhibo Liu, Yuanyuan Yuan, Shuai Wang, Xiaofei Xie, Lei Ma

Research Collection School Of Computing and Information Systems

Due to their widespread use on heterogeneous hardware devices, deep learning (DL) models are compiled into executables by DL compilers to fully leverage low-level hardware primitives. This approach allows DL computations to be undertaken at low cost across a variety of computing platforms, including CPUs, GPUs, and various hardware accelerators. We present BTD (Bin to DNN), a decompiler for deep neural network (DNN) executables. BTD takes DNN executables and outputs full model specifications, including types of DNN operators, network topology, dimensions, and parameters that are (nearly) identical to those of the input models. BTD delivers a practical framework to process …