In this technical report, we present Skywork-13B, a family of large language models (LLMs) trained on a corpus of over 3.2 trillion tokens drawn from both English and Chinese texts. This bilingual foundation model is the most extensively trained and openly published LLMs of comparable size to date. We introduce a two-stage training methodology using a segmented corpus, targeting general purpose training and then domain-specific enhancement training, respectively. We show that our model not only excels on popular benchmarks, but also achieves \emph{state of the art} performance in Chinese language modeling on diverse domains. Furthermore, we propose a novel leakage detection method, demonstrating that test data contamination is a pressing issue warranting further investigation by the LLM community. To spur future research, we release Skywork-13B along with checkpoints obtained during intermediate stages of the training process. We are also releasing part of our SkyPile corpus, a collection of over 150 billion tokens of web text, which is the largest high quality open Chinese pre-training corpus to date. We hope Skywork-13B and our open corpus will serve as a valuable open-source resource to democratize access to high-quality LLMs.
Autonomous agriculture applications (e.g., inspection, phenotyping, plucking fruits) require manipulating the plant foliage to look behind the leaves and the branches. Partial visibility, extreme clutter, thin structures, and unknown geometry and dynamics for plants make such manipulation challenging. We tackle these challenges through data-driven methods. We use self-supervision to train SRPNet, a neural network that predicts what space is revealed on execution of a candidate action on a given plant. We use SRPNet with the cross-entropy method to predict actions that are effective at revealing space beneath plant foliage. Furthermore, as SRPNet does not just predict how much space is revealed but also where it is revealed, we can execute a sequence of actions that incrementally reveal more and more space beneath the plant foliage. We experiment with a synthetic (vines) and a real plant (Dracaena) on a physical test-bed across 5 settings including 2 settings that test generalization to novel plant configurations. Our experiments reveal the effectiveness of our overall method, PPG, over a competitive hand-crafted exploration method, and the effectiveness of SRPNet over a hand-crafted dynamics model and relevant ablations.
The paradigm of pre-training followed by fine-tuning on downstream tasks has become the mainstream method in natural language processing tasks. Although pre-trained models have the advantage of generalization, their performance may still vary significantly across different domain tasks. This is because the data distribution in different domains varies. For example, the different parts of the sentence 'He married Smt. Dipali Ghosh in 1947 and led a very happy married life' may have different impact for downstream tasks. For similarity calculations, words such as 'led' and 'life' are more important. On the other hand, for sentiment analysis, the word 'happy' is crucial. This indicates that different downstream tasks have different levels of sensitivity to sentence components. Our starting point is to scale information of the model and data according to the specifics of downstream tasks, enhancing domain information of relevant parts for these tasks and reducing irrelevant elements for different domain tasks, called SIFTER. In the experimental part, we use the SIFTER to improve SimCSE by constructing positive sample pairs based on enhancing the sentence stem and reducing the unimportant components in the sentence, and maximize the similarity between three sentences. Similarly, SIFTER can improve the gate mechanism of the LSTM model by short-circuiting the input gate of important words so that the LSTM model remembers the important parts of the sentence. Our experiments demonstrate that SIFTER outperforms the SimCSE and LSTM baselines.
Deep learning techniques have been widely used in computed tomography (CT) but require large data sets to train networks. Moreover, data sharing among multiple institutions is limited due to data privacy constraints, which hinders the development of high-performance DL-based CT imaging models from multi-institutional collaborations. Federated learning (FL) strategy is an alternative way to train the models without centralizing data from multi-institutions. In this work, we propose a novel peer-to-peer federated continual learning strategy to improve low-dose CT imaging performance from multiple institutions. The newly proposed method is called peer-to-peer continual FL with intermediate controllers, i.e., icP2P-FL. Specifically, different from the conventional FL model, the proposed icP2P-FL does not require a central server that coordinates training information for a global model. In the proposed icP2P-FL method, the peer-to-peer federated continual learning is introduced wherein the DL-based model is continually trained one client after another via model transferring and inter institutional parameter sharing due to the common characteristics of CT data among the clients. Furthermore, an intermediate controller is developed to make the overall training more flexible. Numerous experiments were conducted on the AAPM low-dose CT Grand Challenge dataset and local datasets, and the experimental results showed that the proposed icP2P-FL method outperforms the other comparative methods both qualitatively and quantitatively, and reaches an accuracy similar to a model trained with pooling data from all the institutions.
With more and more deep neural networks being deployed as various daily services, their reliability is essential. It's frightening that deep neural networks are vulnerable and sensitive to adversarial attacks, the most common one of which for the services is evasion-based. Recent works usually strengthen the robustness by adversarial training or leveraging the knowledge of an amount of clean data. However, in practical terms, retraining and redeploying the model need a large computational budget, leading to heavy losses to the online service. In addition, when adversarial examples of a certain attack are detected, only limited adversarial examples are available for the service provider, while much clean data may not be accessible. Given the mentioned problems, we propose a new scenario, RaPiD (Rapid Plug-in Defender), which is to rapidly defend against a certain attack for the frozen original service model with limitations of few clean and adversarial examples. Motivated by the generalization and the universal computation ability of pre-trained transformer models, we come up with a new defender method, CeTaD, which stands for Considering Pre-trained Transformers as Defenders. In particular, we evaluate the effectiveness and the transferability of CeTaD in the case of one-shot adversarial examples and explore the impact of different parts of CeTaD as well as training data conditions. CeTaD is flexible, able to be embedded into an arbitrary differentiable model, and suitable for various types of attacks.
Deep models have demonstrated recent success in single-image dehazing. Most prior methods consider fully supervised training and learn from paired clean and hazy images, where a hazy image is synthesized based on a clean image and its estimated depth map. This paradigm, however, can produce low-quality hazy images due to inaccurate depth estimation, resulting in poor generalization of the trained models. In this paper, we explore an alternative approach for generating paired clean-hazy images by leveraging computer graphics. Using a modern game engine, our approach renders crisp clean images and their precise depth maps, based on which high-quality hazy images can be synthesized for training dehazing models. To this end, we present SimHaze: a new synthetic haze dataset. More importantly, we show that training with SimHaze alone allows the latest dehazing models to achieve significantly better performance in comparison to previous dehazing datasets. Our dataset and code will be made publicly available.
Learning to optimize (L2O) has emerged as a powerful framework for black-box optimization (BBO). L2O learns the optimization strategies from the target task automatically without human intervention. This paper focuses on obtaining better performance when handling high-dimensional and expensive BBO with little function evaluation cost, which is the core challenge of black-box optimization. However, current L2O-based methods are weak for this due to a large number of evaluations on expensive black-box functions during training and poor representation of optimization strategy. To achieve this, 1) we utilize the cheap surrogate functions of the target task to guide the design of the optimization strategies; 2) drawing on the mechanism of evolutionary algorithm (EA), we propose a novel framework called B2Opt, which has a stronger representation of optimization strategies. Compared to the BBO baselines, B2Opt can achieve 3 to $10^6$ times performance improvement with less function evaluation cost. We test our proposal in high-dimensional synthetic functions and two real-world applications. We also find that deep B2Opt performs better than shallow ones.
Localization in a pre-built map is a basic technique for robot autonomous navigation. Existing mapping and localization methods commonly work well in small-scale environments. As a map grows larger, however, more memory is required and localization becomes inefficient. To solve these problems, map sparsification becomes a practical necessity to acquire a subset of the original map for localization. Previous map sparsification methods add a quadratic term in mixed-integer programming to enforce a uniform distribution of selected landmarks, which requires high memory capacity and heavy computation. In this paper, we formulate map sparsification in an efficient linear form and select uniformly distributed landmarks based on 2D discretized grids. Furthermore, to reduce the influence of different spatial distributions between the mapping and query sequences, which is not considered in previous methods, we also introduce a space constraint term based on 3D discretized grids. The exhaustive experiments in different datasets demonstrate the superiority of the proposed methods in both efficiency and localization performance. The relevant codes will be released at https://github.com/fishmarch/SLAM_Map_Compression.
Conversational recommender systems (CRSs) often utilize external knowledge graphs (KGs) to introduce rich semantic information and recommend relevant items through natural language dialogues. However, original KGs employed in existing CRSs are often incomplete and sparse, which limits the reasoning capability in recommendation. Moreover, only few of existing studies exploit the dialogue context to dynamically refine knowledge from KGs for better recommendation. To address the above issues, we propose the Variational Reasoning over Incomplete KGs Conversational Recommender (VRICR). Our key idea is to incorporate the large dialogue corpus naturally accompanied with CRSs to enhance the incomplete KGs; and perform dynamic knowledge reasoning conditioned on the dialogue context. Specifically, we denote the dialogue-specific subgraphs of KGs as latent variables with categorical priors for adaptive knowledge graphs refactor. We propose a variational Bayesian method to approximate posterior distributions over dialogue-specific subgraphs, which not only leverages the dialogue corpus for restructuring missing entity relations but also dynamically selects knowledge based on the dialogue context. Finally, we infuse the dialogue-specific subgraphs to decode the recommendation and responses. We conduct experiments on two benchmark CRSs datasets. Experimental results confirm the effectiveness of our proposed method.
Camera physiological measurement is a fast growing field of computer vision. Remote photoplethysmography (rPPG) uses video cameras (imagers) to measure the peripheral blood volume pulse (BVP). Simply, this enables heart rate measurement via webcams, smartphone cameras and many other imaging devices. The current state-of-the-art methods are supervised deep neural architectures that have large numbers of parameters and a signal number of hyperparameters. Replication of results and benchmarking of new models is critical for scientific progress. However, as with many other applications of deep learning, reliable codebases are not easy to find. We present a comprehensive toolbox, rPPG-Toolbox, containing code for training and evaluating unsupervised and supervised rPPG models: https://github.com/ubicomplab/rPPG-Toolbox