Xinyu Yang

Hanshi Sun, Zhuoming Chen, Xinyu Yang et al.

With large language models (LLMs) widely deployed in long content generation recently, there has emerged an increasing demand for efficient long-sequence inference support. However, key-value (KV) cache, which is stored to avoid re-computation, has emerged as a critical bottleneck by growing linearly in size with the sequence length. Due to the auto-regressive nature of LLMs, the entire KV cache will be loaded for every generated token, resulting in low utilization of computational cores and high latency. While various compression methods for KV cache have been proposed to alleviate this issue, they suffer from degradation in generation quality. We introduce TriForce, a hierarchical speculative decoding system that is scalable for long sequence generation. This approach leverages the original model weights and dynamic sparse KV cache via retrieval as a draft model, which serves as an intermediate layer in the hierarchy and is further speculated by a smaller model to reduce its drafting latency. TriForce not only facilitates impressive speedups for Llama2-7B-128K, achieving up to 2.31$\times$ on an A100 GPU but also showcases scalability in handling even longer contexts. For the offloading setting on two RTX 4090 GPUs, TriForce achieves 0.108s/token$\unicode{x2014}$only half as slow as the auto-regressive baseline on an A100, which attains 7.78$\times$ on our optimized offloading system. Additionally, TriForce performs 4.86$\times$ than DeepSpeed-Zero-Inference on a single RTX 4090 GPU. TriForce's robustness is highlighted by its consistently outstanding performance across various temperatures. The code is available at https://github.com/Infini-AI-Lab/TriForce.

Harry Dong, Xinyu Yang, Zhenyu Zhang et al.

Many computational factors limit broader deployment of large language models. In this paper, we focus on a memory bottleneck imposed by the key-value (KV) cache, a computational shortcut that requires storing previous KV pairs during decoding. While existing KV cache methods approach this problem by pruning or evicting large swaths of relatively less important KV pairs to dramatically reduce the memory footprint of the cache, they can have limited success in tasks that require recollecting a majority of previous tokens. To alleviate this issue, we propose LESS, a simple integration of a (nearly free) constant sized cache with eviction-based cache methods, such that all tokens can be queried at later decoding steps. Its ability to retain information throughout time shows merit on a variety of tasks where we demonstrate LESS can help reduce the performance gap from caching everything, sometimes even matching it, all while being efficient. Relevant code can be found at https://github.com/hdong920/LESS.

Chaoxiong Ma, Yan Liang, Xinyu Yang et al.

The theory of evidence reasoning has been applied to collective decision-making in recent years. However, existing distributed evidence fusion methods lead to participants' preference leakage and fusion failures as they directly exchange raw evidence and do not assess evidence credibility like centralized credible evidence fusion (CCEF) does. To do so, a privacy-preserving distributed credible evidence fusion method with three-level consensus (PCEF) is proposed in this paper. In evidence difference measure (EDM) neighbor consensus, an evidence-free equivalent expression of EDM among neighbored agents is derived with the shared dot product protocol for pignistic probability and the identical judgment of two events with maximal subjective probabilities, so that evidence privacy is guaranteed due to such irreversible evidence transformation. In EDM network consensus, the non-neighbored EDMs are inferred and neighbored EDMs reach uniformity via interaction between linear average consensus (LAC) and low-rank matrix completion with rank adaptation to guarantee EDM consensus convergence and no solution of inferring raw evidence in numerical iteration style. In fusion network consensus, a privacy-preserving LAC with a self-cancelling differential privacy term is proposed, where each agent adds its randomness to the sharing content and step-by-step cancels such randomness in consensus iterations. Besides, the sufficient condition of the convergence to the CCEF is explored, and it is proven that raw evidence is impossibly inferred in such an iterative consensus. The simulations show that PCEF is close to CCEF both in credibility and fusion results and obtains higher decision accuracy with less time-comsuming than existing methods.

Xinyu Yang, Xinlan Zhang, Zhenguo Zhang et al.

Access to labeled time series data is often limited in the real world, which constrains the performance of deep learning models in the field of time series analysis. Data augmentation is an effective way to solve the problem of small sample size and imbalance in time series datasets. The two key factors of data augmentation are the distance metric and the choice of interpolation method. SMOTE does not perform well on time series data because it uses a Euclidean distance metric and interpolates directly on the object. Therefore, we propose a DTW-based synthetic minority oversampling technique using siamese encoder for interpolation named DTWSSE. In order to reasonably measure the distance of the time series, DTW, which has been verified to be an effective method forts, is employed as the distance metric. To adapt the DTW metric, we use an autoencoder trained in an unsupervised self-training manner for interpolation. The encoder is a Siamese Neural Network for mapping the time series data from the DTW hidden space to the Euclidean deep feature space, and the decoder is used to map the deep feature space back to the DTW hidden space. We validate the proposed methods on a number of different balanced or unbalanced time series datasets. Experimental results show that the proposed method can lead to better performance of the downstream deep learning model.