The dual-encoder has become the de facto architecture for dense retrieval. Typically, it computes the latent representations of the query and document independently, thus failing to fully capture the interactions between the query and document. To alleviate this, recent research has focused on obtaining query-informed document representations. During training, it expands the document with a real query, but during inference, it replaces the real query with a generated one. This inconsistency between training and inference causes the dense retrieval model to prioritize query information while disregarding the document when computing the document representation. Consequently, it performs even worse than the vanilla dense retrieval model because its performance heavily relies on the relevance between the generated queries and the real query.In this paper, we propose a curriculum sampling strategy that utilizes pseudo queries during training and progressively enhances the relevance between the generated query and the real query. By doing so, the retrieval model learns to extend its attention from the document alone to both the document and query, resulting in high-quality query-informed document representations. Experimental results on both in-domain and out-of-domain datasets demonstrate that our approach outperforms previous dense retrieval models.
相關內容
Due to the omnipresence of Neural Radiance Fields (NeRFs), the interest towards editable implicit 3D representations has surged over the last years. However, editing implicit or hybrid representations as used for NeRFs is difficult due to the entanglement of appearance and geometry encoded in the model parameters. Despite these challenges, recent research has shown first promising steps towards photorealistic and non-photorealistic appearance edits. The main open issues of related work include limited interactivity, a lack of support for local edits and large memory requirements, rendering them less useful in practice. We address these limitations with LAENeRF, a unified framework for photorealistic and non-photorealistic appearance editing of NeRFs. To tackle local editing, we leverage a voxel grid as starting point for region selection. We learn a mapping from expected ray terminations to final output color, which can optionally be supervised by a style loss, resulting in a framework which can perform photorealistic and non-photorealistic appearance editing of selected regions. Relying on a single point per ray for our mapping, we limit memory requirements and enable fast optimization. To guarantee interactivity, we compose the output color using a set of learned, modifiable base colors, composed with additive layer mixing. Compared to concurrent work, LAENeRF enables recoloring and stylization while keeping processing time low. Furthermore, we demonstrate that our approach surpasses baseline methods both quantitatively and qualitatively.
Neural implicit scene representations have recently shown encouraging results in dense visual SLAM. However, existing methods produce low-quality scene reconstruction and low-accuracy localization performance when scaling up to large indoor scenes and long sequences. These limitations are mainly due to their single, global radiance field with finite capacity, which does not adapt to large scenarios. Their end-to-end pose networks are also not robust enough with the growth of cumulative errors in large scenes. To this end, we present PLGSLAM, a neural visual SLAM system which performs high-fidelity surface reconstruction and robust camera tracking in real time. To handle large-scale indoor scenes, PLGSLAM proposes a progressive scene representation method which dynamically allocates new local scene representation trained with frames within a local sliding window. This allows us to scale up to larger indoor scenes and improves robustness (even under pose drifts). In local scene representation, PLGSLAM utilizes tri-planes for local high-frequency features. We also incorporate multi-layer perceptron (MLP) networks for the low-frequency feature, smoothness, and scene completion in unobserved areas. Moreover, we propose local-to-global bundle adjustment method with a global keyframe database to address the increased pose drifts on long sequences. Experimental results demonstrate that PLGSLAM achieves state-of-the-art scene reconstruction results and tracking performance across various datasets and scenarios (both in small and large-scale indoor environments). The code will be open-sourced upon paper acceptance.
Neural Radiance Fields (NeRF) have demonstrated impressive potential in synthesizing novel views from dense input, however, their effectiveness is challenged when dealing with sparse input. Existing approaches that incorporate additional depth or semantic supervision can alleviate this issue to an extent. However, the process of supervision collection is not only costly but also potentially inaccurate, leading to poor performance and generalization ability in diverse scenarios. In our work, we introduce a novel model: the Collaborative Neural Radiance Fields (ColNeRF) designed to work with sparse input. The collaboration in ColNeRF includes both the cooperation between sparse input images and the cooperation between the output of the neural radiation field. Through this, we construct a novel collaborative module that aligns information from various views and meanwhile imposes self-supervised constraints to ensure multi-view consistency in both geometry and appearance. A Collaborative Cross-View Volume Integration module (CCVI) is proposed to capture complex occlusions and implicitly infer the spatial location of objects. Moreover, we introduce self-supervision of target rays projected in multiple directions to ensure geometric and color consistency in adjacent regions. Benefiting from the collaboration at the input and output ends, ColNeRF is capable of capturing richer and more generalized scene representation, thereby facilitating higher-quality results of the novel view synthesis. Extensive experiments demonstrate that ColNeRF outperforms state-of-the-art sparse input generalizable NeRF methods. Furthermore, our approach exhibits superiority in fine-tuning towards adapting to new scenes, achieving competitive performance compared to per-scene optimized NeRF-based methods while significantly reducing computational costs. Our code is available at: //github.com/eezkni/ColNeRF.
To segment a signal into blocks to be analyzed, few-shot keyword spotting (KWS) systems often utilize a sliding window of fixed size. Because of the varying lengths of different keywords or their spoken instances, choosing the right window size is a problem: A window should be long enough to contain all necessary information needed to recognize a keyword but a longer window may contain irrelevant information such as multiple words or noise and thus makes it difficult to reliably detect on- and offsets of keywords. We propose TACos, a novel angular margin loss for deriving two-dimensional embeddings that retain temporal properties of the underlying speech signal. In experiments conducted on KWS-DailyTalk, a few-shot KWS dataset presented in this work, using these embeddings as templates for dynamic time warping is shown to outperform using other representations or a sliding window and that using time-reversed segments of the keywords during training improves the performance.
Unveiling Diversity: Empowering OSS Project Leaders with Community Diversity and Turnover Dashboards
Managing open-source software (OSS) projects requires managing communities of contributors. In particular, it is essential for project leaders to understand their community's diversity and turnover. We present CommunityTapestry, a dynamic real-time community dashboard, which presents key diversity and turnover signals that we identified from the literature and through participatory design sessions with stakeholders. We evaluated CommunityTapestry with an OSS project's contributors and Project Management Committee members, who explored the dashboard using their own project data. Our study results demonstrate that CommunityTapestry increased participants' awareness of their community composition and the diversity and turnover rates in the project. It helped them identify areas of improvement and gave them actionable information.
Traditional robotic systems require complex implementations that are not always accessible or easy to use for Human-Robot Interaction (HRI) application developers. With the aim of simplifying the implementation of HRI applications, this paper introduces a novel real-time operating system (RTOS) designed for customizable HRI - RoboSync. By creating multi-level abstraction layers, the system enables users to define complex emotional and behavioral models without needing deep technical expertise. The system's modular architecture comprises a behavior modeling layer, a machine learning plugin configuration layer, a sensor checks customization layer, a scheduler that fits the need of HRI, and a communication and synchronization layer. This approach not only promotes ease of use without highly specialized skills but also ensures real-time responsiveness and adaptability. The primary functionality of the RTOS has been implemented for proof of concept and was tested on a CortexM4 microcontroller, demonstrating its potential for a wide range of lightweight simple-to-implement social robotics applications.
Model editing techniques modify a minor proportion of knowledge in Large Language Models (LLMs) at a relatively low cost, which have demonstrated notable success. Existing methods assume Transformer Layer (TL) hidden states are values of key-value memories of the Feed-Forward Network (FFN). They usually optimize the TL hidden states to memorize target knowledge and use it to update the weights of the FFN in LLMs. However, the information flow of TL hidden states comes from three parts: Multi-Head Self-Attention (MHSA), FFN, and residual connections. Existing methods neglect the fact that the TL hidden states contains information not specifically required for FFN. Consequently, the performance of model editing decreases. To achieve more precise model editing, we analyze hidden states of MHSA and FFN, finding that MHSA encodes certain general knowledge extraction patterns. This implies that MHSA weights do not require updating when new knowledge is introduced. Based on above findings, we introduce PMET, which simultaneously optimizes Transformer Component (TC, namely MHSA and FFN) hidden states, while only using the optimized TC hidden states of FFN to precisely update FFN weights. Our experiments demonstrate that PMET exhibits state-of-the-art performance on both the COUNTERFACT and zsRE datasets. Our ablation experiments substantiate the effectiveness of our enhancements, further reinforcing the finding that the MHSA encodes certain general knowledge extraction patterns and indicating its storage of a small amount of factual knowledge. Our code is available at //github.com/xpq-tech/PMET.
Prompt recently have become an effective linguistic tool on utilizing the pre-trained language models. However, in few-shot scenarios, subtle changes of prompt's design always make the result widely different, and the prompt design is also easy to overfit the current limited samples. To alleviate this, we explore how to utilize suitable contrastive samples and multiple contrastive learning methods to realize a more robust prompt's representation. Therefore, the contrastive prompt model ConsPrompt combining with prompt encoding network, contrastive sampling modules, and contrastive scoring modules are introduced to realize differential contrastive learning. Our results exhibit the state-of-the-art performance in different few-shot settings, and the ablation experiments also certificate the effectiveness in utilizing multi-degree contrastive learning in prompt-based fine-tuning process.
The costly self-attention layers in modern Transformers require memory and compute quadratic in sequence length. Existing approximation methods usually underperform and fail to obtain significant speedups in practice. Here we present SwitchHead - a novel method that reduces both compute and memory requirements and achieves wall-clock speedup, while matching the language modeling performance of baseline Transformers with the same parameter budget. SwitchHead uses Mixture-of-Experts (MoE) layers for the value and output projections and requires 4 to 8 times fewer attention matrices than standard Transformers. Our novel attention can also be combined with MoE MLP layers, resulting in an efficient fully-MoE "SwitchHead" Transformer model. Our code is public.
A timely software update is vital to combat the increasing security vulnerabilities. However, some software vendors may secretly patch their vulnerabilities without creating CVE entries or even describing the security issue in their change log. Thus, it is critical to identify these hidden security patches and defeat potential N-day attacks. Researchers have employed various machine learning techniques to identify security patches in open-source software, leveraging the syntax and semantic features of the software changes and commit messages. However, all these solutions cannot be directly applied to the binary code, whose instructions and program flow may dramatically vary due to different compilation configurations. In this paper, we propose BinGo, a new security patch detection system for binary code. The main idea is to present the binary code as code property graphs to enable a comprehensive understanding of program flow and perform a language model over each basic block of binary code to catch the instruction semantics. BinGo consists of four phases, namely, patch data pre-processing, graph extraction, embedding generation, and graph representation learning. Due to the lack of an existing binary security patch dataset, we construct such a dataset by compiling the pre-patch and post-patch source code of the Linux kernel. Our experimental results show BinGo can achieve up to 80.77% accuracy in identifying security patches between two neighboring versions of binary code. Moreover, BinGo can effectively reduce the false positives and false negatives caused by the different compilers and optimization levels.
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191