Search

In distributed fiber-optic sensing based on optical frequency domain reflectometry (OFDR), Doppler frequency shifts due to the changes of disturbances during one sweep period introduce demodulation errors that accumulate along both the distance and time, impairing the sensing performance. Here, we report distributed dynamic strain sensing using Doppler-shift-immune phase-sensitive OFDR based on frequency-tracking and spectrum-zooming with ultra-weak reflection array. Theoretical study has been carried out with the introduction of mismatch coefficient, unveiling quantitatively the impact of Doppler shift. Following a numerical analysis of the proposed method, a retained precision has been experimentally verified regardless of the position mismatch due to the Doppler effect. Doppler-shift-immune sensing for dynamic strains covering continuous spatial resolution over a distance of 1000 m with a 2.5 cm sensing spatial resolution has been demonstrated, verifying the high fidelity promised by the proposed method.

We examine diverging preferences in human-labeled preference datasets. We develop a taxonomy of disagreement sources spanning 10 categories across four high-level classes -- task underspecification, response style, refusals, and annotation errors. We find that the majority of disagreements are in opposition with standard reward modeling approaches, which are designed with the assumption that annotator disagreement is noise. We then explore how these findings impact two areas of LLM development: reward modeling and evaluation. In our experiments, we demonstrate how standard reward modeling methods, like the Bradley-Terry model, fail to differentiate whether a given preference judgment is the result of unanimous agreement among annotators or the majority opinion among diverging user preferences. We also find that these tendencies are also echoed by popular LLM-as-Judge evaluation methods, which consistently identify a winning response in cases of diverging preferences. These findings highlight remaining challenges in LLM evaluations, which are greatly influenced by divisive features like response style, and in developing pluralistically aligned LLMs. To address these issues, we develop methods for identifying diverging preferences to mitigate their influence on evaluation and training.

Reward models are critical for aligning models to follow instructions, and are typically trained following one of two popular paradigms: Bradley-Terry style or Regression style. However, there is a lack of evidence that either approach is better than the other, when adequately matched for data. This is primarily because these approaches require data collected in different (but incompatible) formats, meaning that adequately matched data is not available in existing public datasets. To tackle this problem, we release preference annotations (designed for Bradley-Terry training) to complement existing ratings (designed for Regression style training) in the HelpSteer2 dataset. To improve data interpretability, preference annotations are accompanied with human-written justifications. Using this data, we conduct the first head-to-head comparison of Bradley-Terry and Regression models when adequately matched for data. Based on insights derived from such a comparison, we propose a novel approach to combine Bradley-Terry and Regression reward modeling. A Llama-3.1-70B-Instruct model tuned with this approach scores 94.1 on RewardBench, emerging top of more than 140 reward models as of 1 Oct 2024. We also demonstrate the effectiveness of this reward model at aligning models to follow instructions in RLHF. We open-source this dataset (CC-BY-4.0 license) at https://huggingface.co/datasets/nvidia/HelpSteer2 and openly release the trained Reward Model at https://huggingface.co/nvidia/Llama-3.1-Nemotron-70B-Reward, Comment: 26 pages, 3 figures

Guardrails have become an integral part of Large language models (LLMs), by moderating harmful or toxic response in order to maintain LLMs' alignment to human expectations. However, the existing guardrail methods do not consider different needs and access rights of individual users, and treat all the users with the same rule. This study introduces an adaptive guardrail mechanism, supported by trust modeling and enhanced with in-context learning, to dynamically modulate access to sensitive content based on user trust metrics. By leveraging a combination of direct interaction trust and authority-verified trust, the system precisely tailors the strictness of content moderation to align with the user's credibility and the specific context of their inquiries. Our empirical evaluations demonstrate that the adaptive guardrail effectively meets diverse user needs, outperforming existing guardrails in practicality while securing sensitive information and precisely managing potentially hazardous content through a context-aware knowledge base. This work is the first to introduce trust-oriented concept within a guardrail system, offering a scalable solution that enriches the discourse on ethical deployment for next-generation LLMs., Comment: Under Review

Websites are critical in today's digital world, with over 1.11 billion currently active and approximately 252,000 new sites launched daily. Converting website layout design into functional UI code is a time-consuming yet indispensable step of website development. Manual methods of converting visual designs into functional code present significant challenges, especially for non-experts. To explore automatic design-to-code solutions, we first conduct a motivating study on GPT-4o and identify three types of issues in generating UI code: element omission, element distortion, and element misarrangement. We further reveal that a focus on smaller visual segments can help multimodal large language models (MLLMs) mitigate these failures in the generation process. In this paper, we propose DCGen, a divide-and-conquer-based approach to automate the translation of webpage design to UI code. DCGen starts by dividing screenshots into manageable segments, generating descriptions for each segment, and then reassembling them into complete UI code for the entire screenshot. We conduct extensive testing with a dataset comprised of real-world websites and various MLLMs and demonstrate that DCGen achieves up to a 14% improvement in visual similarity over competing methods. To the best of our knowledge, DCGen is the first segment-aware prompt-based approach for generating UI code directly from screenshots.

We release the Nemotron-4 340B model family, including Nemotron-4-340B-Base, Nemotron-4-340B-Instruct, and Nemotron-4-340B-Reward. Our models are open access under the NVIDIA Open Model License Agreement, a permissive model license that allows distribution, modification, and use of the models and its outputs. These models perform competitively to open access models on a wide range of evaluation benchmarks, and were sized to fit on a single DGX H100 with 8 GPUs when deployed in FP8 precision. We believe that the community can benefit from these models in various research studies and commercial applications, especially for generating synthetic data to train smaller language models. Notably, over 98% of data used in our model alignment process is synthetically generated, showcasing the effectiveness of these models in generating synthetic data. To further support open research and facilitate model development, we are also open-sourcing the synthetic data generation pipeline used in our model alignment process.

High-quality preference datasets are essential for training reward models that can effectively guide large language models (LLMs) in generating high-quality responses aligned with human preferences. As LLMs become stronger and better aligned, permissively licensed preference datasets, such as Open Assistant, HH-RLHF, and HelpSteer need to be updated to remain effective for reward modeling. Methods that distil preference data from proprietary LLMs such as GPT-4 have restrictions on commercial usage imposed by model providers. To improve upon both generated responses and attribute labeling quality, we release HelpSteer2, a permissively licensed preference dataset (CC-BY-4.0). Using a powerful internal base model trained on HelpSteer2, we are able to achieve the SOTA score (92.0%) on Reward-Bench's primary dataset, outperforming currently listed open and proprietary models, as of June 12th, 2024. Notably, HelpSteer2 consists of only ten thousand response pairs, an order of magnitude fewer than existing preference datasets (e.g., HH-RLHF), which makes it highly efficient for training reward models. Our extensive experiments demonstrate that reward models trained with HelpSteer2 are effective in aligning LLMs. In particular, we propose SteerLM 2.0, a model alignment approach that can effectively make use of the rich multi-attribute score predicted by our reward models. HelpSteer2 is available at https://huggingface.co/datasets/nvidia/HelpSteer2 and code is available at https://github.com/NVIDIA/NeMo-Aligner

The China Space Station Telescope (CSST) is a two-meter space telescope with multiple back-end instruments. The Fine Guidance Sensor (FGS) is an essential subsystem of the CSST Precision Image Stability System to ensure the required absolute pointing accuracy and line-of-sight stabilization. In this study, we construct the Main Guide Star Catalog for FGS. To accomplish this, we utilize the information about the FGS and object information from the Gaia Data Release 3. We provide an FGS instrument magnitude and exclude variables, binaries, and high proper motion stars from the catalog to ensure uniform FGS guidance capabilities. Subsequently, we generate a HEALPix index, which provides a hierarchical tessellation of the celestial sphere, and employ the Voronoi algorithm to achieve a homogeneous distribution of stars across the catalog. This distribution ensures adequate coverage and sampling of the sky. The performance of the CSST guide star catalog was assessed by simulating the field of view of the FGS according to the CSST mock survey strategy catalog. The analysis of the results indicates that this catalog provides adequate coverage and accuracy. The catalog's performance meets the FGS requirements, ensuring the functioning of the FGS and its guidance capabilities., Comment: published on RAA

In the burgeoning field of Large Language Models (LLMs), developing a robust safety mechanism, colloquially known as "safeguards" or "guardrails", has become imperative to ensure the ethical use of LLMs within prescribed boundaries. This article provides a systematic literature review on the current status of this critical mechanism. It discusses its major challenges and how it can be enhanced into a comprehensive mechanism dealing with ethical issues in various contexts. First, the paper elucidates the current landscape of safeguarding mechanisms that major LLM service providers and the open-source community employ. This is followed by the techniques to evaluate, analyze, and enhance some (un)desirable properties that a guardrail might want to enforce, such as hallucinations, fairness, privacy, and so on. Based on them, we review techniques to circumvent these controls (i.e., attacks), to defend the attacks, and to reinforce the guardrails. While the techniques mentioned above represent the current status and the active research trends, we also discuss several challenges that cannot be easily dealt with by the methods and present our vision on how to implement a comprehensive guardrail through the full consideration of multi-disciplinary approach, neural-symbolic method, and systems development lifecycle., Comment: under review. arXiv admin note: text overlap with arXiv:2402.01822

The remarkable generative capabilities of diffusion models have motivated extensive research in both image and video editing. Compared to video editing which faces additional challenges in the time dimension, image editing has witnessed the development of more diverse, high-quality approaches and more capable software like Photoshop. In light of this gap, we introduce a novel and generic solution that extends the applicability of image editing tools to videos by propagating edits from a single frame to the entire video using a pre-trained image-to-video model. Our method, dubbed I2VEdit, adaptively preserves the visual and motion integrity of the source video depending on the extent of the edits, effectively handling global edits, local edits, and moderate shape changes, which existing methods cannot fully achieve. At the core of our method are two main processes: Coarse Motion Extraction to align basic motion patterns with the original video, and Appearance Refinement for precise adjustments using fine-grained attention matching. We also incorporate a skip-interval strategy to mitigate quality degradation from auto-regressive generation across multiple video clips. Experimental results demonstrate our framework's superior performance in fine-grained video editing, proving its capability to produce high-quality, temporally consistent outputs., Comment: 19 pages

Aligning Large Language Models (LLMs) with human values and preferences is essential for making them helpful and safe. However, building efficient tools to perform alignment can be challenging, especially for the largest and most competent LLMs which often contain tens or hundreds of billions of parameters. We create NeMo-Aligner, a toolkit for model alignment that can efficiently scale to a thousand GPUs for training the largest open-source LLMs such as Nemotron 4 340B and Llama 3.1 405B. NeMo-Aligner comes with highly optimized and scalable implementations for major paradigms of model alignment such as: Reinforcement Learning from Human Feedback (RLHF), Direct Preference Optimization (DPO), SteerLM, and Self-Play Fine-Tuning (SPIN). Additionally, our toolkit supports running most of the alignment techniques in a Parameter Efficient Fine-Tuning (PEFT) setting. NeMo-Aligner is designed for extensibility, allowing support for other alignment techniques with minimal effort. It is open-sourced with Apache 2.0 License and we invite community contributions at https://github.com/NVIDIA/NeMo-Aligner, Comment: 16 pages, 4 figures, Accepted to COLM 2024

As Large Language Models (LLMs) become more integrated into our daily lives, it is crucial to identify and mitigate their risks, especially when the risks can have profound impacts on human users and societies. Guardrails, which filter the inputs or outputs of LLMs, have emerged as a core safeguarding technology. This position paper takes a deep look at current open-source solutions (Llama Guard, Nvidia NeMo, Guardrails AI), and discusses the challenges and the road towards building more complete solutions. Drawing on robust evidence from previous research, we advocate for a systematic approach to construct guardrails for LLMs, based on comprehensive consideration of diverse contexts across various LLMs applications. We propose employing socio-technical methods through collaboration with a multi-disciplinary team to pinpoint precise technical requirements, exploring advanced neural-symbolic implementations to embrace the complexity of the requirements, and developing verification and testing to ensure the utmost quality of the final product., Comment: Proceedings of the 41st International Conference on Machine Learning, Vienna, Austria. PMLR 235, 2024

In the realm of power systems, the increasing involvement of residential users in load forecasting applications has heightened concerns about data privacy. Specifically, the load data can inadvertently reveal the daily routines of residential users, thereby posing a risk to their property security. While federated learning (FL) has been employed to safeguard user privacy by enabling model training without the exchange of raw data, these FL models have shown vulnerabilities to emerging attack techniques, such as Deep Leakage from Gradients and poisoning attacks. To counteract these, we initially employ a Secure-Aggregation (SecAgg) algorithm that leverages multiparty computation cryptographic techniques to mitigate the risk of gradient leakage. However, the introduction of SecAgg necessitates the deployment of additional sub-center servers for executing the multiparty computation protocol, thereby escalating computational complexity and reducing system robustness, especially in scenarios where one or more sub-centers are unavailable. To address these challenges, we introduce a Markovian Switching-based distributed training framework, the convergence of which is substantiated through rigorous theoretical analysis. The Distributed Markovian Switching (DMS) topology shows strong robustness towards the poisoning attacks as well. Case studies employing real-world power system load data validate the efficacy of our proposed algorithm. It not only significantly minimizes communication complexity but also maintains accuracy levels comparable to traditional FL methods, thereby enhancing the scalability of our load forecasting algorithm.

The optimisation of crop harvesting processes for commonly cultivated crops is of great importance in the aim of agricultural industrialisation. Nowadays, the utilisation of machine vision has enabled the automated identification of crops, leading to the enhancement of harvesting efficiency, but challenges still exist. This study presents a new framework that combines two separate architectures of convolutional neural networks (CNNs) in order to simultaneously accomplish the tasks of crop detection and harvesting (robotic manipulation) inside a simulated environment. Crop images in the simulated environment are subjected to random rotations, cropping, brightness, and contrast adjustments to create augmented images for dataset generation. The you only look once algorithmic framework is employed with traditional rectangular bounding boxes for crop localization. The proposed method subsequently utilises the acquired image data via a visual geometry group model in order to reveal the grasping positions for the robotic manipulation., Comment: 7 pages, 9 figures

This paper considers the distributed leader-follower stress-matrix-based affine formation control problem of discrete-time linear multi-agent systems with static and dynamic leaders. In leader-follower multi-agent formation control, the aim is to drive a set of agents comprising leaders and followers to form any desired geometric pattern and simultaneously execute any required manoeuvre by controlling only a few agents denoted as leaders. Existing works in literature are mostly limited to the cases where the agents' inter-agent communications are either in the continuous-time settings or the sampled-data cases where the leaders are constrained to constant (or zero) velocities or accelerations. Here, we relax these constraints and study the discrete-time cases where the leaders can have stationary or time-varying velocities. We propose control laws in the study of different situations and provide some sufficient conditions to guarantee the overall system stability. Simulation study is used to demonstrate the efficacy of our proposed control laws.

Image collage is a very useful tool for visualizing an image collection. Most of the existing methods and commercial applications for generating image collages are designed on simple shapes, such as rectangular and circular layouts. This greatly limits the use of image collages in some artistic and creative settings. Although there are some methods that can generate irregularly-shaped image collages, they often suffer from severe image overlapping and excessive blank space. This prevents such methods from being effective information communication tools. In this paper, we present a shape slicing algorithm and an optimization scheme that can create image collages of arbitrary shapes in an informative and visually pleasing manner given an input shape and an image collection. To overcome the challenge of irregular shapes, we propose a novel algorithm, called Shape-Aware Slicing, which partitions the input shape into cells based on medial axis and binary slicing tree. Shape-Aware Slicing, which is designed specifically for irregular shapes, takes human perception and shape structure into account to generate visually pleasing partitions. Then, the layout is optimized by analyzing input images with the goal of maximizing the total salient regions of the images. To evaluate our method, we conduct extensive experiments and compare our results against previous work. The evaluations show that our proposed algorithm can efficiently arrange image collections on irregular shapes and create visually superior results than prior work and existing commercial tools., Comment: This paper has been accepted for publication on IEEE Transactions on Visualization and Computer Graphics (TVCG), March 2023. Project website http://graphics.csie.ncku.edu.tw/shapedimagecollage

Existing open-source helpfulness preference datasets do not specify what makes some responses more helpful and others less so. Models trained on these datasets can incidentally learn to model dataset artifacts (e.g. preferring longer but unhelpful responses only due to their length). To alleviate this problem, we collect HelpSteer, a multi-attribute helpfulness dataset annotated for the various aspects that make responses helpful. Specifically, our 37k-sample dataset has annotations for correctness, coherence, complexity, and verbosity in addition to overall helpfulness of responses. Training Llama 2 70B using the HelpSteer dataset with SteerLM technique produces a model that scores 7.54 on MT Bench, which is currently the highest score for open models that do not require training data from more powerful models (e.g. GPT4). We release this dataset with CC-BY-4.0 license at https://huggingface.co/datasets/nvidia/HelpSteer

Model alignment with human preferences is an essential step in making Large Language Models (LLMs) helpful and consistent with human values. It typically consists of supervised fine-tuning (SFT) and reinforcement learning from human feedback (RLHF) stages. However, RLHF faces inherent limitations stemming from a complex training setup and its tendency to align the model with implicit values that end users cannot control at run-time. Moreover, reward models in RLHF stage commonly rely on single-dimensional feedback as opposed to explicit, multifaceted signals that indicate attributes such as helpfulness, humor, and toxicity. To address these limitations, we propose SteerLM, a supervised fine-tuning method that empowers end-users to control responses during inference. SteerLM conditions responses to conform to an explicitly defined multi-dimensional set of attributes, thereby empowering a steerable AI capable of generating helpful and high-quality responses while maintaining customizability. Experiments show that SteerLM trained on open source datasets generates responses that are preferred by human and automatic evaluators to many state-of-the-art baselines trained with RLHF while being much easier to train. Try SteerLM at https://huggingface.co/nvidia/SteerLM-llama2-13B, Comment: Findings of EMNLP 2023

The potential of the muon colliders open up new possibilities for the exploration of new physics beyond the Standard Model. It is worthwhile to investigate whether muon colliders are suitable for studying gluonic quartic gauge couplings~(gQGCs), which can be contributed by dimension-8 operators in the framework of the Standard Model effective field theory, and are intensively studied recently. In this paper, we study the sensitivity of the process $\mu^+\mu^-\to j j \nu\bar{\nu}$ to gQGCs. Our result indicate that the muon colliders with c.m. energies larger than $4\;{\rm TeV}$ can be more sensitive to gQGCs than the Large Hadron Collider., Comment: 20 pages, 5 figures