Starting in 2003 when the first MWE workshop was held with ACL in Sapporo, Japan, this year, the joint workshop of MWE-UD co-located with the LREC-COLING 2024 conference marked the 20th anniversary of MWE workshop events over the past nearly two decades. Standing at this milestone, we look back to this workshop series and summarise the research topics and methodologies researchers have carried out over the years. We also discuss the current challenges that we are facing and the broader impacts/synergies of MWE research within the CL and NLP fields. Finally, we give future research perspectives. We hope this position paper can help researchers, students, and industrial practitioners interested in MWE get a brief but easy understanding of its history, current, and possible future.
Collaborative recommendation fundamentally involves learning high-quality user and item representations from interaction data. Recently, graph convolution networks (GCNs) have advanced the field by utilizing high-order connectivity patterns in interaction graphs, as evidenced by state-of-the-art methods like PinSage and LightGCN. However, one key limitation has not been well addressed in existing solutions: capturing long-range collaborative filtering signals, which are crucial for modeling user preference. In this work, we propose a new graph transformer (GT) framework -- \textit{Position-aware Graph Transformer for Recommendation} (PGTR), which combines the global modeling capability of Transformer blocks with the local neighborhood feature extraction of GCNs. The key insight is to explicitly incorporate node position and structure information from the user-item interaction graph into GT architecture via several purpose-designed positional encodings. The long-range collaborative signals from the Transformer block are then combined linearly with the local neighborhood features from the GCN backbone to enhance node embeddings for final recommendations. Empirical studies demonstrate the effectiveness of the proposed PGTR method when implemented on various GCN-based backbones across four real-world datasets, and the robustness against interaction sparsity as well as noise.
This paper presents a robust position controller for electric power assisted steering and steer-by-wire force-feedback systems. A position controller is required in steering systems for haptic feedback control, advanced driver assistance systems and automated driving. However, the driver's \textit{physical} arm impedance causes an inertial uncertainty during coupling. Consequently, a typical position controller, i.e., based on single variable, becomes less robust and suffers tracking performance loss. Therefore, a robust position controller is investigated. The proposed solution is based on the multi-variable concept such that the sensed driver torque signal is also included in the position controller. The subsequent solution is obtained by solving the LMI$-H_{\infty}$ optimization problem. As a result, the desired loop gain shape is achieved, i.e., large gain at low frequencies for performance and small gain at high frequencies for robustness. Finally, frequency response comparison of different position controllers on real hardware is presented. Experiments and simulation results clearly illustrate the improvements in reference tracking and robustness with the proposed $H_\infty$ controller.
Vision-centric autonomous driving systems require diverse data for robust training and evaluation, which can be augmented by manipulating object positions and appearances within existing scene captures. While recent advancements in diffusion models have shown promise in video editing, their application to object manipulation in driving scenarios remains challenging due to imprecise positional control and difficulties in preserving high-fidelity object appearances. To address these challenges in position and appearance control, we introduce DriveEditor, a diffusion-based framework for object editing in driving videos. DriveEditor offers a unified framework for comprehensive object editing operations, including repositioning, replacement, deletion, and insertion. These diverse manipulations are all achieved through a shared set of varying inputs, processed by identical position control and appearance maintenance modules. The position control module projects the given 3D bounding box while preserving depth information and hierarchically injects it into the diffusion process, enabling precise control over object position and orientation. The appearance maintenance module preserves consistent attributes with a single reference image by employing a three-tiered approach: low-level detail preservation, high-level semantic maintenance, and the integration of 3D priors from a novel view synthesis model. Extensive qualitative and quantitative evaluations on the nuScenes dataset demonstrate DriveEditor's exceptional fidelity and controllability in generating diverse driving scene edits, as well as its remarkable ability to facilitate downstream tasks. Project page: https://yvanliang.github.io/DriveEditor.
This paper studies a sub-connected six-dimensional movable antenna (6DMA)-aided multi-user communication system. In this system, each sub-array is connected to a dedicated radio frequency chain and collectively moves and rotates as a unit within specific local regions. The movement and rotation capabilities of 6DMAs enhance design flexibility, facilitating the capture of spatial variations for improved communication performance. To fully characterize the effect of antenna position and orientation on wireless channels between the base station (BS) and users, we develop a field-response-based 6DMA channel model to account for the antenna radiation pattern and polarization. We then maximize the sum rate of multiple users, by jointly optimizing the digital and unit-modulus analog beamformers given the transmit power budget as well as the positions and orientations of sub-arrays within given movable and rotatable ranges at the BS. Due to the highly coupled variables, the formulated optimization problem is non-convex and thus challenging to solve. We develop a fractional programming-aided alternating optimization framework that integrates the Lagrange multiplier method, manifold optimization, and gradient descent to solve the problem. Numerical results demonstrate that the proposed 6DMA-aided sub-connected structure achieves a substantial sum-rate improvement over various benchmark schemes with less flexibility in antenna movement and can even outperform fully-digital beamforming systems that employ antenna position or orientation adjustments only. The results also highlight the necessity of considering antenna polarization for optimally adjusting antenna orientation.
Gastrointestinal (GI) bleeding, a critical indicator of digestive system disorders, re quires efficient and accurate detection methods. This paper presents our solution to the Auto-WCEBleedGen Version V1 Challenge, where we achieved the consolation position. We developed a unified YOLOv8-X model for both detection and classification of bleeding regions in Wireless Capsule Endoscopy (WCE) images. Our approach achieved 96.10% classification accuracy and 76.8% mean Average Precision (mAP) at 0.5 IoU on the val idation dataset. Through careful dataset curation and annotation, we assembled and trained on 6,345 diverse images to ensure robust model performance. Our implementa tion code and trained models are publicly available at https://github.com/pavan98765/Auto-WCEBleedGen.
Extending the context length of Language Models (LMs) by improving Rotary Position Embedding (RoPE) has become a trend. While existing works mainly address RoPE's limitations within attention mechanism, this paper provides an analysis across nearly all parts of LMs, uncovering their adverse effects on length generalization for RoPE-based attention. Using Discrete Signal Processing theory, we show that RoPE enables periodic attention by implicitly achieving Non-Uniform Discrete Fourier Transform. However, this periodicity is undermined by the spectral damage caused by: 1) linear layers and activation functions outside of attention; 2) insufficiently trained frequency components brought by time-domain truncation. Building on our observations, we propose Fourier Position Embedding (FoPE), which enhances attention's frequency-domain properties to improve both its periodic extension and length generalization. FoPE constructs Fourier Series and zero-outs the destructive frequency components, increasing model robustness against the spectrum damage. Experiments across various model scales and benchmarks show that, within varying context windows, FoPE maintains a more stable performance compared to RoPE and ALiBi. Several analyses and ablations bring further support to our method and theoretical modeling.
Embedding models are crucial for tasks in Information Retrieval (IR) and semantic similarity measurement, yet their handling of longer texts and associated positional biases remains underexplored. In this study, we investigate the impact of content position and input size on text embeddings. Our experiments reveal that embedding models, irrespective of their positional encoding mechanisms, disproportionately prioritize the beginning of an input. Ablation studies demonstrate that insertion of irrelevant text or removal at the start of a document reduces cosine similarity between altered and original embeddings by up to 12.3% more than ablations at the end. Regression analysis further confirms this bias, with sentence importance declining as position moves further from the start, even with with content-agnosticity. We hypothesize that this effect arises from pre-processing strategies and chosen positional encoding techniques. These findings quantify the sensitivity of retrieval systems and suggest a new lens towards embedding model robustness.
This position paper investigates the potential of integrating insights from language impairment research and its clinical treatment to develop human-inspired learning strategies and evaluation frameworks for language models (LMs). We inspect the theoretical underpinnings underlying some influential linguistically motivated training approaches derived from neurolinguistics and, particularly, aphasiology, aimed at enhancing the recovery and generalization of linguistic skills in aphasia treatment, with a primary focus on those targeting the syntactic domain. We highlight how these insights can inform the design of rigorous assessments for LMs, specifically in their handling of complex syntactic phenomena, as well as their implications for developing human-like learning strategies, aligning with efforts to create more sustainable and cognitively plausible natural language processing (NLP) models.
Large language models (LLMs) have made significant advancements across various tasks, but their safety alignment remain a major concern. Exploring jailbreak prompts can expose LLMs' vulnerabilities and guide efforts to secure them. Existing methods primarily design sophisticated instructions for the LLM to follow, or rely on multiple iterations, which could hinder the performance and efficiency of jailbreaks. In this work, we propose a novel jailbreak paradigm, Simple Assistive Task Linkage (SATA), which can effectively circumvent LLM safeguards and elicit harmful responses. Specifically, SATA first masks harmful keywords within a malicious query to generate a relatively benign query containing one or multiple [MASK] special tokens. It then employs a simple assistive task such as a masked language model task or an element lookup by position task to encode the semantics of the masked keywords. Finally, SATA links the assistive task with the masked query to jointly perform the jailbreak. Extensive experiments show that SATA achieves state-of-the-art performance and outperforms baselines by a large margin. Specifically, on AdvBench dataset, with mask language model (MLM) assistive task, SATA achieves an overall attack success rate (ASR) of 85% and harmful score (HS) of 4.57, and with element lookup by position (ELP) assistive task, SATA attains an overall ASR of 76% and HS of 4.43.
In the software industry, the drive to add new features often overshadows the need to improve existing code. Large Language Models (LLMs) offer a new approach to improving codebases at an unprecedented scale through AI-assisted refactoring. However, LLMs come with inherent risks such as braking changes and the introduction of security vulnerabilities. We advocate for encapsulating the interaction with the models in IDEs and validating refactoring attempts using trustworthy safeguards. However, equally important for the uptake of AI refactoring is research on trust development. In this position paper, we position our future work based on established models from research on human factors in automation. We outline action research within CodeScene on development of 1) novel LLM safeguards and 2) user interaction that conveys an appropriate level of trust. The industry collaboration enables large-scale repository analysis and A/B testing to continuously guide the design of our research interventions.
With the adoption of retrieval-augmented generation (RAG), large language models (LLMs) are expected to ground their generation to the retrieved contexts. Yet, this is hindered by position bias of LLMs, failing to evenly attend to all contexts. Previous work has addressed this by synthesizing contexts with perturbed positions of gold segment, creating a position-diversified train set. We extend this intuition to propose consistency regularization with augmentation and distillation. First, we augment each training instance with its position perturbation to encourage consistent predictions, regardless of ordering. We also distill behaviors of this pair, although it can be counterproductive in certain RAG scenarios where the given order from the retriever is crucial for generation quality. We thus propose CORD, balancing COnsistency and Rank Distillation. CORD adaptively samples noise-controlled perturbations from an interpolation space, ensuring both consistency and respect for the rank prior. Empirical results show this balance enables CORD to outperform consistently in diverse RAG benchmarks.
Recent years have witnessed rapid development in the research area of WiFi sensing, which senses human activities in a contactless and non-intrusive manner. One major issue that hinders real-world deployment of these systems is position dependence, i.e., once the human target changes location and orientation, the sensing performance degrades significantly. Existing machine learning based methods aim to solve this problem by either generating high-dimensional features or transfer learning the environment knowledge. However, these methods require significant training effort and yet acquire limited improvement. In this paper, we start by understanding and analyzing the Doppler frequency shift in WiFi sensing. We then develop a WiFi frequency model to quantify the relationship between signal frequency and target position, motion direction and speed for human activities. Based on this theoretical model, we prove that the commonly-used movement speed and motion direction features are position dependent, and further identify movement fragments and relative motion direction changes as two position-independent features. Building upon the frequency model and the position-independent features, we design a suite of position-independent gestures and develop the gesture recognition system accordingly. Evaluation results show that under various conditions (i.e., different locations, orientations, environments, and persons), our system achieves more than 96 percent recognition accuracy without any training, significantly outperforming state-of-the-art machine learning based solutions.
Recent advancements in view synthesis have significantly enhanced immersive experiences across various computer graphics and multimedia applications, including telepresence, and entertainment. By enabling the generation of new perspectives from a single input view, view synthesis allows users to better perceive and interact with their environment. However, many state-of-the-art methods, while achieving high visual quality, face limitations in real-time performance, which makes them less suitable for live applications where low latency is critical. In this paper, we present a lightweight, position-aware network designed for real-time view synthesis from a single input image and a target camera pose. The proposed framework consists of a Position Aware Embedding, modeled with a multi-layer perceptron, which efficiently maps positional information from the target pose to generate high dimensional feature maps. These feature maps, along with the input image, are fed into a Rendering Network that merges features from dual encoder branches to resolve both high level semantics and low level details, producing a realistic new view of the scene. Experimental results demonstrate that our method achieves superior efficiency and visual quality compared to existing approaches, particularly in handling complex translational movements without explicit geometric operations like warping. This work marks a step toward enabling real-time view synthesis from a single image for live and interactive applications.
Theory of Mind (ToM) capabilities in LLMs have recently become a central object of investigation. Cognitive science distinguishes between two steps required for ToM tasks: 1) determine whether to invoke ToM, which includes the appropriate Depth of Mentalizing (DoM), or level of recursion required to complete a task; and 2) applying the correct inference given the DoM. In this position paper, we first identify several lines of work in different communities in AI, including LLM benchmarking, ToM add-ons, ToM probing, and formal models for ToM. We argue that recent work in AI tends to focus exclusively on the second step which are typically framed as static logic problems. We conclude with suggestions for improved evaluation of ToM capabilities inspired by dynamic environments used in cognitive tasks.
In the realm of gradient-based optimization, Nesterov's accelerated gradient method (NAG) is a landmark advancement, achieving an accelerated convergence rate that outperforms the vanilla gradient descent method for convex function. However, for strongly convex functions, whether NAG converges linearly remains an open question, as noted in the comprehensive review by Chambolle and Pock [2016]. This issue, aside from the critical step size, was addressed by Li et al. [2024a] using a high-resolution differential equation framework. Furthermore, Beck [2017, Section 10.7.4] introduced a monotonically convergent variant of NAG, referred to as M-NAG. Despite these developments, the Lyapunov analysis presented in [Li et al., 2024a] cannot be directly extended to M-NAG. In this paper, we propose a modification to the iterative relation by introducing a gradient term, leading to a new gradient-based iterative relation. This adjustment allows for the construction of a novel Lyapunov function that excludes kinetic energy. The linear convergence derived from this Lyapunov function is independent of both the parameters of the strongly convex functions and the step size, yielding a more general and robust result. Notably, we observe that the gradient iterative relation derived from M-NAG is equivalent to that from NAG when the position-velocity relation is applied. However, the Lyapunov analysis does not rely on the position-velocity relation, allowing us to extend the linear convergence to M-NAG. Finally, by utilizing two proximal inequalities, which serve as the proximal counterparts of strongly convex inequalities, we extend the linear convergence to both the fast iterative shrinkage-thresholding algorithm (FISTA) and its monotonic counterpart (M-FISTA).
Human readers can accurately count how many letters are in a word (e.g., 7 in ``buffalo''), remove a letter from a given position (e.g., ``bufflo'') or add a new one. The human brain of readers must have therefore learned to disentangle information related to the position of a letter and its identity. Such disentanglement is necessary for the compositional, unbounded, ability of humans to create and parse new strings, with any combination of letters appearing in any positions. Do modern deep neural models also possess this crucial compositional ability? Here, we tested whether neural models that achieve state-of-the-art on disentanglement of features in visual input can also disentangle letter position and letter identity when trained on images of written words. Specifically, we trained beta variational autoencoder ($\beta$-VAE) to reconstruct images of letter strings and evaluated their disentanglement performance using CompOrth - a new benchmark that we created for studying compositional learning and zero-shot generalization in visual models for orthography. The benchmark suggests a set of tests, of increasing complexity, to evaluate the degree of disentanglement between orthographic features of written words in deep neural models. Using CompOrth, we conducted a set of experiments to analyze the generalization ability of these models, in particular, to unseen word length and to unseen combinations of letter identities and letter positions. We found that while models effectively disentangle surface features, such as horizontal and vertical `retinal' locations of words within an image, they dramatically fail to disentangle letter position and letter identity and lack any notion of word length. Together, this study demonstrates the shortcomings of state-of-the-art $\beta$-VAE models compared to humans and proposes a new challenge and a corresponding benchmark to evaluate neural models.
The effective utilization of unlicensed spectrum is regarded as an important direction to enable the massive access and broad coverage for next-generation wireless local area network (WLAN). Due to the crowded spectrum occupancy and dense user terminals (UTs), the conventional fixed antenna (FA)-based access points (APs) face huge challenges in realizing massive access and interference cancellation. To address this issue, in this paper we develop a six-dimensional movable antenna (6DMA) enhanced multi-AP coordination system for coverage enhancement and interference mitigation. First, we model the wireless channels between the APs and UTs to characterize their variation with respect to 6DMA movement, in terms of both the three-dimensional (3D) position and 3D orientation of each distributed AP's antenna. Then, an optimization problem is formulated to maximize the weighted sum rate of multiple UTs for their uplink transmissions by jointly optimizing the antenna position vector (APV), the antenna orientation matrix (AOM), and the receive combining matrix over all coordinated APs, subject to the constraints on local antenna movement regions. To solve this challenging non-convex optimization problem, we first transform it into a more tractable Lagrangian dual problem. Then, an alternating optimization (AO)-based algorithm is developed by iteratively optimizing the APV and AOM, which are designed by applying the successive convex approximation (SCA) technique and Riemannian manifold optimization-based algorithm, respectively. Simulation results show that the proposed 6DMA-enhanced multi-AP coordination system can significantly enhance network capacity, and both of the online and offline 6DMA schemes can attain considerable performance improvement compared to the conventional FA-based schemes.
We consider a variant of the target defense problems where a group of defenders are tasked to simultaneously capture an intruder. The intruder's objective is to reach a target without being simultaneously captured by the defender team. Some of the defenders are sensing-limited and do not have any information regarding the intruder's position or velocity at any time. The defenders may communicate with each other using a connected communication graph. We propose a decentralized feedback strategy for the defenders, which transforms the simultaneous capture problem into a unique nonlinear consensus problem. We derive a sufficient condition for simultaneous capture in terms of the agents' speeds, sensing, and communication capabilities. The proposed decentralized controller is evaluated through extensive numerical simulations.
In order to promote agricultural automatic picking and yield estimation technology, this project designs a set of automatic detection, positioning and counting algorithms for grape bunches, and applies it to agricultural robots. The Yolov3 detection network is used to realize the accurate detection of grape bunches, and the local tracking algorithm is added to eliminate relocation. Then it obtains the accurate 3D spatial position of the central points of grape bunches using the depth distance and the spatial restriction method. Finally, the counting of grape bunches is completed. It is verified using the agricultural robot in the simulated vineyard environment. The project code is released at: https://github.com/XuminGaoGithub/Grape_bunches_count_using_robots.