Jialun Cao, Yaojie Lu, Meiziniu Li, Haoyang Ma, Haokun Li, Mengda He, Cheng Wen, Le Sun, Hongyu Zhang, Shengchao Qin, Shing-Chi Cheung, Cong Tian. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2025.

The COVID-19 pandemic, with its unprecedented challenges and disruptions, triggered a profound transformation in the retail industry . Health and safety regulations including periodic lockdowns , supply chain disruptions, and economic uncertainty affected the way businesses operate and led to a drastic change in consumer behaviour. This triggered the modification of traditional sales trends, which in turn impacted the accuracy of existing demand forecasting methods, and will affect their future performance. Moreover, the reliance of machine learning algorithms on historical sales data for training made them ill-equipped to adapt to these abrupt sales pattern shifts. Therefore, innovative solutions to address the complexities arising from this new landscape are needed. This paper introduces a framework aimed at enhancing demand forecasting accuracy in the post-pandemic period. Central to this framework is a feature engineering approach involving the creation of a predictor variable that encapsulates the level of restrictions imposed during the pandemic across seven distinct categories. This granular approach not only accounts for lockdowns and store closures but also considers indirect factors influencing retail sales, such as remote work arrangements and school closures. An extensive empirical evaluation of the proposed approach was conducted on a real-world retail dataset obtained from Charles Clinkard, a UK-based footwear retailer, demonstrating consistent improvements in forecasting accuracy across four deep probabilistic models and three levels of product aggregation in a real-life setting. Further validation utilising the Retail Sales Index dataset, which reflects monthly sales across various retail sectors in Great Britain, was also undertaken using six forecasting models, corroborating our initial findings. Overall, we show that leveraging historical sales data spanning the pandemic period – or, in general, any period where the data has inherent bias – is still viable for training machine learning models to forecast the demand, provided that an effective feature engineering approach is implemented.

Abstract Formal verification provides a rigorous and systematic approach to ensure the correctness and reliability of software systems. Yet, constructing specifications for the full proof relies on domain expertise and non-trivial manpower. In view of such needs, an automated approach for specification synthesis is desired. While existing automated approaches are limited in their versatility, i.e. , they either focus only on synthesizing loop invariants for numerical programs, or are tailored for specific types of programs or invariants. Programs involving multiple complicated data types ( e.g. , arrays, pointers) and code structures ( e.g. , nested loops, function calls) are often beyond their capabilities. To help bridge this gap, we present AutoSpec , an automated approach to synthesize specifications for automated program verification. It overcomes the shortcomings of existing work in specification versatility, synthesizing satisfiable and adequate specifications for full proof. It is driven by static analysis and program verification, and is empowered by large language models (LLMs). AutoSpec addresses the practical challenges in three ways: (1) driving AutoSpec by static analysis and program verification, LLMs serve as generators to generate candidate specifications, (2) programs are decomposed to direct the attention of LLMs, and (3) candidate specifications are validated in each round to avoid error accumulation during the interaction with LLMs. In this way, AutoSpec can incrementally and iteratively generate satisfiable and adequate specifications. The evaluation shows its effectiveness and usefulness, as it outperforms existing works by successfully verifying 79% of programs through automatic specification synthesis, a significant improvement of 1.592x. It can also be successfully applied to verify the programs in a real-world X509-parser project.

Rust libraries are ubiquitous in Rust-based software development. Guaranteeing their correctness and reliability requires thorough analysis and testing. Fuzzing is a popular bug-finding solution, yet it requires writing fuzz targets for libraries. Recently, some automatic fuzz target generation methods have been proposed. However, two challenges remain: (1) how to generate diverse API sequences that prioritize unsafe code and interactions to reveal bugs in Rust libraries; (2) how to provide support for the generic APIs and verify both syntactic and semantic validity of the fuzz targets to enable more comprehensive testing of Rust libraries. In this paper, we propose RPG, an automatic fuzz target synthesis technique to support Rust library fuzzing. RPG uses a pool-based search to generate diverse and unsafe API sequences, and synthesizes fuzz targets with generic support and validity check. The experimental results demonstrate that RPG enhances both the quality of the generated fuzz targets and the bug-finding ability through pool-based generation and generic support, substantially outperforming the state-of-the-art. Moreover, RPG has discovered 25 previously unknown bugs from 50 well-known Rust libraries available on Crates.io.

The C and C++ languages introduced the relaxed-memory concurrency into the language specification for efficiency purposes in 2011. Trace semantics can provide the mathematical foundation for the proposed C++11 memory model, and there is a lack of investigation of trace semantics for C++11. The Promising Semantics (PS) of Kang et al. provides the standard SC-style operational semantics for the C++11 concurrency model, where “SC” refers to “Sequential Consistency”. Inspired by PS, in this article we first investigate the trace semantics for the relaxed read and write accesses under C++11, acting in the denotational semantics style. In our semantic model, a trace is in the form of a sequence of snapshots, and the snapshots record the modification in the relevant global or local variables, and the thread view. Moreover, the trace semantics for the release/acquire accesses under C++11 is also explored, based on the separated thread views and newly added message views. When considering this trace model, different accesses bring in their unique snapshots, and make distinguished effects on the production of the sequences. For any given program, the proposed trace semantics in this article produces all the valid traces directly. Furthermore, our trace semantics, together with that for TSO and MCA ARMv8, has the possibility to be the foundation of the meta model of the trace semantics for weak memory models.

The C++11 standard introduced a language level weak memory model (i.e., the C++11 memory model) to improve the performance of the execution of C/C++ programs. Algebra is well-suited for direct use by engineers in symbolic calculation of parameters. It is a challenge to investigate the algebraic semantics for the C++11 memory model. Inspired by the promising semantics, in this paper, we explore the algebraic laws for the C++11 memory model, including a set of sequential and parallel expansion laws. We introduce the concept of guarded choice, and every program under the C++11 memory model can be converted into the head normal form of guarded choice. In addition, the proposed algebraic laws are implemented in the rewriting engine Maude.

Controlled concurrency testing (CCT) techniques have been shown promising for concurrency bug detection. Their key insight is to control the order in which threads get executed, and attempt to explore the space of possible interleavings of a concurrent program to detect bugs. However, various challenges remain in current CCT techniques, rendering them ineffective and ad-hoc. In this paper, we propose a novel CCT technique Period. Unlike previous works, Period models the execution of concurrent programs as periodical execution, and systematically explores the space of possible inter-leavings, where the exploration is guided by periodical scheduling and influenced by previously tested interleavings. We have evaluated Period on 10 real-world CVEs and 36 widely-used benchmark programs, and our experimental results show that Period demonstrates superiority over other CCT techniques in both effectiveness and runtime overhead. Moreover, we have discovered 5 previously unknown concurrency bugs in real-world programs.

In this paper, we revisit model predictive control (MPC) for the classical wheeled mobile robot (WMR) navigation problem. We prove that the reachable set based hierarchical MPC (HMPC), a state-of-the-art MPC, cannot handle WMR navigation in theory due to the non-existence of non-trivial linear system with an under-approximate reachable set of WMR. Nevertheless, we propose a virtual linear leader guided MPC (VLL-MPC) to enable HMPC structure. Different from current HMPCs, we use a virtual linear system with an under-approximate path set rather than the traditional trace set to guide the WMR. We provide a valid construction of the virtual linear leader. We prove the stability of VLL-MPC, and discuss its complexity. In the experiment, we demonstrate the advantage of VLL-MPC empirically by comparing it with NMPC, LMPC and anytime RRT* in several scenarios.

We show how complex privacy requirements can be represented and processed by an extended model of Attribute Based Access Control (ABAC), working with a simple database applications pattern. During application model development, most likely based on UML (e.g. Use Case, Class Diagrams), the analyst and possibly the end user specifies ABAC permissions, and then verifies their effect by running queries on the target data. The ABAC model supports positive and negative permissions, "break glass" overrides of negative permissions, and message/alert generation. The permissions combining algorithms are based on relational database optimisation, and permissions processing is implemented by query modification, producing structurally-optimised queries in an SQL-like language; the queries can then be processed by many database and big data systems. The method and models have been implemented in a prototype Privacy Preferences Tool in collaboration with a large medical records development, and we discuss experiences with focus group evaluations of this tool.

By accepting order weakening for memory operations, the C11 memory model allows C/C++ programs to take advantage of modern hardware architectures, where weak/relaxed memory models are now the norm. However, the weakened C11 memory model introduces many complex and counterintuitive behaviours, rendering it more difficult for people to understand or reason about concurrent C11 programs. Several program logics (RSL, GPS, FSL, GPS+) have been proposed over the last few years to support formal reasoning for C11 programs, but each of them deals with only a specific subset of C11 programs, mainly due to the high complexity of the weakened memory model. Notably none of these program logics supports the reasoning of release-sequences - a highly flexible synchronisation mechanism in C11. Very recently, Doko and Vafeiadis propose a way in their FSL++ logic to reason about C11 programs using release-sequences, but their solution is restricted to those scenarios where only atomic update operations are between the release head and the receiver. In this paper we propose a new program logic that offers full support for reasoning about C11 programs using release-sequences. Our proposed logic is built on top of our previous program logic GPS+, but with much finer control over the resource transmission by introducing restricted-shareable assertions and an enhanced protocol system. We also illustrate our approach by verifying release-sequence programs that existing logics would not be able to.

Automatic facial expression recognition plays an important role in various application domains such as medical imaging, surveillance and human-robot interaction. This research proposes a novel facial expression recognition system with modified Local Gabor Binary Patterns (LGBP) for feature extraction and a firefly algorithm (FA) variant for feature optimization. First of all, in order to deal with illumination changes, scaling differences and rotation variations, we propose an extended overlap LGBP to extract initial discriminative facial features. Then a modified FA is proposed to reduce the dimensionality of the extracted facial features. This FA variant employs Gaussian, Cauchy and Levy distributions to further mutate the best solution identified by the FA to increase exploration in the search space to avoid premature convergence. The overall system is evaluated using three facial expression databases (i.e. CK+, MMI, and JAFFE). The proposed system outperforms other heuristic search algorithms such as Genetic Algorithm and Particle Swarm Optimization and other existing state-of-the-art facial expression recognition research, significantly.

For efficiency reasons, weak (or relaxed) memory is now the norm on modern architectures. To cater for this trend, modern programming languages are adapting their memory models. The new C11 memory model [1] allows several levels of memory weakening, including non-atomics, relaxed atomics, release-acquire atomics, and sequentially consistent atomics. Under such weak memory models, multithreaded programs exhibit more behaviours, some of which would have been inconsistent under the traditional strong (i.e. sequentially consistent) memory model. This makes the task of reasoning about concurrent programs even more challenging. The GPS framework, recently developed by Turon et al.[22], has made a step forward towards tackling this challenge. By integrating ghost states, per-location protocols and separation logic, GPS can successfully verify programs with release-acquire atomics. In this paper, we present a program logic, an enhancement of the GPS framework, that can support the verification of a bigger class of C11 programs, that is, programs with release-acquire atomics, relaxed atomics and release-acquire fences. Key elements of our proposed logic include two new types of assertions, a more expressive resource model and a set of newly-designed verification rules.