MoDULA: Mixture of Domain-Specific and Universal LoRA for Multi-Task Learning

The growing demand for larger-scale models in the development of \textbf{L}arge \textbf{L}anguage \textbf{M}odels (LLMs) poses challenges for efficient training within limited computational resources. Traditional fine-tuning methods often exhibit instability in multi-task learning and rely heavily on extensive training resources. Here, we propose MoDULA (\textbf{M}ixture \textbf{o}f \textbf{D}omain-Specific and \textbf{U}niversal \textbf{L}oR\textbf{A}), a novel \textbf{P}arameter \textbf{E}fficient \textbf{F}ine-\textbf{T}uning (PEFT) \textbf{M}ixture-\textbf{o}f-\textbf{E}xpert (MoE) paradigm for improved fine-tuning and parameter efficiency in multi-task learning. The paradigm effectively improves the multi-task capability of the model by training universal experts, domain-specific experts, and routers separately. MoDULA-Res is a new method within the MoDULA paradigm, which maintains the model's general capability by connecting universal and task-specific experts through residual connections. The experimental results demonstrate that the overall performance of the MoDULA-Flan and MoDULA-Res methods surpasses that of existing fine-tuning methods on various LLMs. Notably, MoDULA-Res achieves more significant performance improvements in multiple tasks while reducing training costs by over 80\% without losing general capability. Moreover, MoDULA displays flexible pluggability, allowing for the efficient addition of new tasks without retraining existing experts from scratch. This progressive training paradigm circumvents data balancing issues, enhancing training efficiency and model stability. Overall, MoDULA provides a scalable, cost-effective solution for fine-tuning LLMs with enhanced parameter efficiency and generalization capability.

AG-SLAM: Active Gaussian Splatting SLAM

We present AG-SLAM, the first active SLAM system utilizing 3D Gaussian Splatting (3DGS) for online scene reconstruction. In recent years, radiance field scene representations, including 3DGS have been widely used in SLAM and exploration, but actively planning trajectories for robotic exploration is still unvisited. In particular, many exploration methods assume precise localization and thus do not mitigate the significant risk of constructing a trajectory, which is difficult for a SLAM system to operate on. This can cause camera tracking failure and lead to failures in real-world robotic applications. Our method leverages Fisher Information to balance the dual objectives of maximizing the information gain for the environment while minimizing the cost of localization errors. Experiments conducted on the Gibson and Habitat-Matterport 3D datasets demonstrate state-of-the-art results of the proposed method.

Next Best Sense: Guiding Vision and Touch with FisherRF for 3D Gaussian Splatting

We propose a framework for active next best view and touch selection for robotic manipulators using 3D Gaussian Splatting (3DGS). 3DGS is emerging as a useful explicit 3D scene representation for robotics, as it has the ability to represent scenes in a both photorealistic and geometrically accurate manner. However, in real-world, online robotic scenes where the number of views is limited given efficiency requirements, random view selection for 3DGS becomes impractical as views are often overlapping and redundant. We address this issue by proposing an end-to-end online training and active view selection pipeline, which enhances the performance of 3DGS in few-view robotics settings. We first elevate the performance of few-shot 3DGS with a novel semantic depth alignment method using Segment Anything Model 2 (SAM2) that we supplement with Pearson depth and surface normal loss to improve color and depth reconstruction of real-world scenes. We then extend FisherRF, a next-best-view selection method for 3DGS, to select views and touch poses based on depth uncertainty. We perform online view selection on a real robot system during live 3DGS training. We motivate our improvements to few-shot GS scenes, and extend depth-based FisherRF to them, where we demonstrate both qualitative and quantitative improvements on challenging robot scenes. For more information, please see our project page at https://armlabstanford.github.io/next-best-sense.

Contrastive Token Learning with Similarity Decay for Repetition Suppression in Machine Translation

For crosslingual conversation and trade, Neural Machine Translation (NMT) is pivotal yet faces persistent challenges with monotony and repetition in generated content. Traditional solutions that rely on penalizing text redundancy or token reoccurrence have shown limited efficacy, particularly for lengthy article and e-commerce descriptions with inherent redundancy, even with the advent of Large Language Models (LLMs). This paper investigates the underlying causes of textual repetition through the lens of information entropy, attributing the phenomenon to the elevated uncertainty within the input text. To address this, a novel algorithm named Contrastive Token Learning with Similarity Decay (CTSD) is introduced, which modulates the suppression of tokens dynamically, informed by varying attention weights and inter-token distances. Furthermore, an e-commerce dataset comprised of title texts of online real items is compiled and released susceptible to hallucination translations to benchmark the algorithm. Extensive evaluations demonstrate that CTSD significantly outperforms existing approaches in precision and generalizability. Additional online A/B testing underscores its practical value, showing marked improvements in user engagement and conversion. Notably, this method has been implemented with full traffic on eight multilingual sites of alibaba.com, the largest B2B e-commerce platform in the world.

Modeling User Intent Beyond Trigger: Incorporating Uncertainty for Trigger-Induced Recommendation

To cater to users' desire for an immersive browsing experience, numerous e-commerce platforms provide various recommendation scenarios, with a focus on Trigger-Induced Recommendation (TIR) tasks. However, the majority of current TIR methods heavily rely on the trigger item to understand user intent, lacking a higher-level exploration and exploitation of user intent (e.g., popular items and complementary items), which may result in an overly convergent understanding of users' short-term intent and can be detrimental to users' long-term purchasing experiences. Moreover, users' short-term intent shows uncertainty and is affected by various factors such as browsing context and historical behaviors, which poses challenges to user intent modeling. To address these challenges, we propose a novel model called Deep Uncertainty Intent Network (DUIN), comprising three essential modules: i) Explicit Intent Exploit Module extracting explicit user intent using the contrastive learning paradigm; ii) Latent Intent Explore Module exploring latent user intent by leveraging the multi-view relationships between items; iii) Intent Uncertainty Measurement Module offering a distributional estimation and capturing the uncertainty associated with user intent. Experiments on three real-world datasets demonstrate the superior performance of DUIN compared to existing baselines. Notably, DUIN has been deployed across all TIR scenarios in our e-commerce platform, with online A/B testing results conclusively validating its superiority.

Sample-agnostic Adversarial Perturbation for Vision-Language Pre-training Models

Recent studies on AI security have highlighted the vulnerability of Vision-Language Pre-training (VLP) models to subtle yet intentionally designed perturbations in images and texts. Investigating multimodal systems' robustness via adversarial attacks is crucial in this field. Most multimodal attacks are sample-specific, generating a unique perturbation for each sample to construct adversarial samples. To the best of our knowledge, it is the first work through multimodal decision boundaries to explore the creation of a universal, sample-agnostic perturbation that applies to any image. Initially, we explore strategies to move sample points beyond the decision boundaries of linear classifiers, refining the algorithm to ensure successful attacks under the top $k$ accuracy metric. Based on this foundation, in visual-language tasks, we treat visual and textual modalities as reciprocal sample points and decision hyperplanes, guiding image embeddings to traverse text-constructed decision boundaries, and vice versa. This iterative process consistently refines a universal perturbation, ultimately identifying a singular direction within the input space which is exploitable to impair the retrieval performance of VLP models. The proposed algorithms support the creation of global perturbations or adversarial patches. Comprehensive experiments validate the effectiveness of our method, showcasing its data, task, and model transferability across various VLP models and datasets. Code: https://github.com/LibertazZ/MUAP

A Unified Understanding of Adversarial Vulnerability Regarding Unimodal Models and Vision-Language Pre-training Models

With Vision-Language Pre-training (VLP) models demonstrating powerful multimodal interaction capabilities, the application scenarios of neural networks are no longer confined to unimodal domains but have expanded to more complex multimodal V+L downstream tasks. The security vulnerabilities of unimodal models have been extensively examined, whereas those of VLP models remain challenging. We note that in CV models, the understanding of images comes from annotated information, while VLP models are designed to learn image representations directly from raw text. Motivated by this discrepancy, we developed the Feature Guidance Attack (FGA), a novel method that uses text representations to direct the perturbation of clean images, resulting in the generation of adversarial images. FGA is orthogonal to many advanced attack strategies in the unimodal domain, facilitating the direct application of rich research findings from the unimodal to the multimodal scenario. By appropriately introducing text attack into FGA, we construct Feature Guidance with Text Attack (FGA-T). Through the interaction of attacking two modalities, FGA-T achieves superior attack effects against VLP models. Moreover, incorporating data augmentation and momentum mechanisms significantly improves the black-box transferability of FGA-T. Our method demonstrates stable and effective attack capabilities across various datasets, downstream tasks, and both black-box and white-box settings, offering a unified baseline for exploring the robustness of VLP models.

Robust Interaction-based Relevance Modeling for Online E-Commerce and LLM-based Retrieval

Semantic relevance calculation is crucial for e-commerce search engines, as it ensures that the items selected closely align with customer intent. Inadequate attention to this aspect can detrimentally affect user experience and engagement. Traditional text-matching techniques are prevalent but often fail to capture the nuances of search intent accurately, so neural networks now have become a preferred solution to processing such complex text matching. Existing methods predominantly employ representation-based architectures, which strike a balance between high traffic capacity and low latency. However, they exhibit significant shortcomings in generalization and robustness when compared to interaction-based architectures. In this work, we introduce a robust interaction-based modeling paradigm to address these shortcomings. It encompasses 1) a dynamic length representation scheme for expedited inference, 2) a professional terms recognition method to identify subjects and core attributes from complex sentence structures, and 3) a contrastive adversarial training protocol to bolster the model's robustness and matching capabilities. Extensive offline evaluations demonstrate the superior robustness and effectiveness of our approach, and online A/B testing confirms its ability to improve relevance in the same exposure position, resulting in more clicks and conversions. To the best of our knowledge, this method is the first interaction-based approach for large e-commerce search relevance calculation. Notably, we have deployed it for the entire search traffic on alibaba.com, the largest B2B e-commerce platform in the world.

Beyond Uncertainty: Risk-Aware Active View Acquisition for Safe Robot Navigation and 3D Scene Understanding with FisherRF

This work proposes a novel approach to bolster both the robot's risk assessment and safety measures while deepening its understanding of 3D scenes, which is achieved by leveraging Radiance Field (RF) models and 3D Gaussian Splatting. To further enhance these capabilities, we incorporate additional sampled views from the environment with the RF model. One of our key contributions is the introduction of Risk-aware Environment Masking (RaEM), which prioritizes crucial information by selecting the next-best-view that maximizes the expected information gain. This targeted approach aims to minimize uncertainties surrounding the robot's path and enhance the safety of its navigation. Our method offers a dual benefit: improved robot safety and increased efficiency in risk-aware 3D scene reconstruction and understanding. Extensive experiments in real-world scenarios demonstrate the effectiveness of our proposed approach, highlighting its potential to establish a robust and safety-focused framework for active robot exploration and 3D scene understanding.

General2Specialized LLMs Translation for E-commerce

Existing Neural Machine Translation (NMT) models mainly handle translation in the general domain, while overlooking domains with special writing formulas, such as e-commerce and legal documents. Taking e-commerce as an example, the texts usually include amounts of domain-related words and have more grammar problems, which leads to inferior performances of current NMT methods. To address these problems, we collect two domain-related resources, including a set of term pairs (aligned Chinese-English bilingual terms) and a parallel corpus annotated for the e-commerce domain. Furthermore, we propose a two-step fine-tuning paradigm (named G2ST) with self-contrastive semantic enhancement to transfer one general NMT model to the specialized NMT model for e-commerce. The paradigm can be used for the NMT models based on Large language models (LLMs). Extensive evaluations on real e-commerce titles demonstrate the superior translation quality and robustness of our G2ST approach, as compared with state-of-the-art NMT models such as LLaMA, Qwen, GPT-3.5, and even GPT-4.