Compute-in-memory (CIM) accelerators built upon non-volatile memory (NVM) devices excel in energy efficiency and latency when performing Deep Neural Network (DNN) inference, thanks to their in-situ data processing capability. However, the stochastic nature and intrinsic variations of NVM devices often result in performance degradation in DNN inference. Introducing these non-ideal device behaviors during DNN training enhances robustness, but drawbacks include limited accuracy improvement, reduced prediction confidence, and convergence issues. This arises from a mismatch between the deterministic training and non-deterministic device variations, as such training, though considering variations, relies solely on the model's final output. In this work, we draw inspiration from the control theory and propose a novel training concept: Negative Feedback Training (NFT) leveraging the multi-scale noisy information captured from network. We develop two specific NFT instances, Oriented Variational Forward (OVF) and Intermediate Representation Snapshot (IRS). Extensive experiments show that our methods outperform existing state-of-the-art methods with up to a 46.71% improvement in inference accuracy while reducing epistemic uncertainty, boosting output confidence, and improving convergence probability. Their effectiveness highlights the generality and practicality of our NFT concept in enhancing DNN robustness against device variations.
相關內容
Recent work in automated program repair (APR) proposes the use of reasoning and patch validation feedback to reduce the semantic gap between the LLMs and the code under analysis. The idea has been shown to perform well for general APR, but its effectiveness in other particular contexts remains underexplored. In this work, we assess the impact of reasoning and patch validation feedback to LLMs in the context of vulnerability repair, an important and challenging task in security. To support the evaluation, we present VRpilot, an LLM-based vulnerability repair technique based on reasoning and patch validation feedback. VRpilot (1) uses a chain-of-thought prompt to reason about a vulnerability prior to generating patch candidates and (2) iteratively refines prompts according to the output of external tools (e.g., compiler, code sanitizers, test suite, etc.) on previously-generated patches. To evaluate performance, we compare VRpilot against the state-of-the-art vulnerability repair techniques for C and Java using public datasets from the literature. Our results show that VRpilot generates, on average, 14% and 7.6% more correct patches than the baseline techniques on C and Java, respectively. We show, through an ablation study, that reasoning and patch validation feedback are critical. We report several lessons from this study and potential directions for advancing LLM-empowered vulnerability repair
Finetuning large language models (LLMs) in federated learning (FL) settings has become important as it allows resource-constrained devices to finetune a model using private data. However, finetuning LLMs using backpropagation requires excessive memory (especially from intermediate activations) for resource-constrained devices. While Forward-mode Auto-Differentiation (AD) can reduce memory footprint from activations, we observe that directly applying it to LLM finetuning results in slow convergence and poor accuracy. This work introduces Spry, an FL algorithm that splits trainable weights of an LLM among participating clients, such that each client computes gradients using Forward-mode AD that are closer estimates of the true gradients. Spry achieves a low memory footprint, high accuracy, and fast convergence. We theoretically show that the global gradients in Spry are unbiased estimates of true global gradients for homogeneous data distributions across clients, while heterogeneity increases bias of the estimates. We also derive Spry's convergence rate, showing that the gradients decrease inversely proportional to the number of FL rounds, indicating the convergence up to the limits of heterogeneity. Empirically, Spry reduces the memory footprint during training by 1.4-7.1$\times$ in contrast to backpropagation, while reaching comparable accuracy, across a wide range of language tasks, models, and FL settings. Spry reduces the convergence time by 1.2-20.3$\times$ and achieves 5.2-13.5\% higher accuracy against state-of-the-art zero-order methods. When finetuning Llama2-7B with LoRA, compared to the peak memory usage of 33.9GB of backpropagation, Spry only consumes 6.2GB of peak memory. For OPT13B, the reduction is from 76.5GB to 10.8GB. Spry makes feasible previously impossible FL deployments on commodity mobile and edge devices. Source code is available at //github.com/Astuary/Spry.
In recent times, large language models (LLMs) have made significant strides in generating computer code, blurring the lines between code created by humans and code produced by artificial intelligence (AI). As these technologies evolve rapidly, it is crucial to explore how they influence code generation, especially given the risk of misuse in areas like higher education. This paper explores this issue by using advanced classification techniques to differentiate between code written by humans and that generated by ChatGPT, a type of LLM. We employ a new approach that combines powerful embedding features (black-box) with supervised learning algorithms - including Deep Neural Networks, Random Forests, and Extreme Gradient Boosting - to achieve this differentiation with an impressive accuracy of 98%. For the successful combinations, we also examine their model calibration, showing that some of the models are extremely well calibrated. Additionally, we present white-box features and an interpretable Bayes classifier to elucidate critical differences between the code sources, enhancing the explainability and transparency of our approach. Both approaches work well but provide at most 85-88% accuracy. We also show that untrained humans solve the same task not better than random guessing. This study is crucial in understanding and mitigating the potential risks associated with using AI in code generation, particularly in the context of higher education, software development, and competitive programming.
Language models (LMs) have long been used to improve results of automatic speech recognition (ASR) systems, but they are unaware of the errors that ASR systems make. Error correction models are designed to fix ASR errors, however, they showed little improvement over traditional LMs mainly due to the lack of supervised training data. In this paper, we present Denoising LM (DLM), which is a $\textit{scaled}$ error correction model trained with vast amounts of synthetic data, significantly exceeding prior attempts meanwhile achieving new state-of-the-art ASR performance. We use text-to-speech (TTS) systems to synthesize audio, which is fed into an ASR system to produce noisy hypotheses, which are then paired with the original texts to train the DLM. DLM has several $\textit{key ingredients}$: (i) up-scaled model and data; (ii) usage of multi-speaker TTS systems; (iii) combination of multiple noise augmentation strategies; and (iv) new decoding techniques. With a Transformer-CTC ASR, DLM achieves 1.5% word error rate (WER) on $\textit{test-clean}$ and 3.3% WER on $\textit{test-other}$ on Librispeech, which to our knowledge are the best reported numbers in the setting where no external audio data are used and even match self-supervised methods which use external audio data. Furthermore, a single DLM is applicable to different ASRs, and greatly surpassing the performance of conventional LM based beam-search rescoring. These results indicate that properly investigated error correction models have the potential to replace conventional LMs, holding the key to a new level of accuracy in ASR systems.
This paper introduces PipeFusion, a novel approach that harnesses multi-GPU parallelism to address the high computational and latency challenges of generating high-resolution images with diffusion transformers (DiT) models. PipeFusion splits images into patches and distributes the network layers across multiple devices. It employs a pipeline parallel manner to orchestrate communication and computations. By leveraging the high similarity between the input from adjacent diffusion steps, PipeFusion eliminates the waiting time in the pipeline by reusing the one-step stale feature maps to provide context for the current step. Our experiments demonstrate that it can generate higher image resolution where existing DiT parallel approaches meet OOM. PipeFusion significantly reduces the required communication bandwidth, enabling DiT inference to be hosted on GPUs connected via PCIe rather than the more costly NVLink infrastructure, which substantially lowers the overall operational expenses for serving DiT models. Our code is publicly available at //github.com/PipeFusion/PipeFusion.
Recent large language models (LLMs) have witnessed significant advancement in various tasks, including mathematical reasoning and theorem proving. As these two tasks require strict and formal multi-step inference, they are appealing domains for exploring the reasoning ability of LLMs but still face important challenges. Previous studies such as Chain-of-Thought (CoT) have revealed the effectiveness of intermediate steps guidance. However, such step-wise annotation requires heavy labor, leading to insufficient training steps for current benchmarks. To fill this gap, this work introduces MUSTARD, a data generation framework that masters uniform synthesis of theorem and proof data of high quality and diversity. MUSTARD synthesizes data in three stages: (1) It samples a few mathematical concept seeds as the problem category. (2) Then, it prompts a generative language model with the sampled concepts to obtain both the problems and their step-wise formal solutions. (3) Lastly, the framework utilizes a proof assistant (e.g., Lean Prover) to filter the valid proofs. With the proposed MUSTARD, we present a theorem-and-proof benchmark MUSTARDSAUCE with 5,866 valid data points. Each data point contains an informal statement, an informal proof, and a translated formal proof that passes the prover validation. We perform extensive analysis and demonstrate that MUSTARD generates validated high-quality step-by-step data. We further apply the MUSTARDSAUCE for fine-tuning smaller language models. The fine-tuned Llama 2-7B achieves a 15.41% average relative performance gain in automated theorem proving, and 8.18% in math word problems. Codes and data are available at //github.com/Eleanor-H/MUSTARD.
eXmY is a novel data type for quantization of ML models. It supports both arbitrary bit widths and arbitrary integer and floating point formats. For example, it seamlessly supports 3, 5, 6, 7, 9 bit formats. For a specific bit width, say 7, it defines all possible formats e.g. e0m6, e1m5, e2m4, e3m3, e4m2, e5m1 and e6m0. For non-power of two bit widths e.g. 5, 6, 7, we created a novel encoding and decoding scheme which achieves perfect compression, byte addressability and is amenable to sharding and vector processing. We implemented libraries for emulation, encoding and decoding tensors and checkpoints in C++, TensorFlow, JAX and PAX. For optimal performance, the codecs use SIMD instructions on CPUs and vector instructions on TPUs and GPUs. eXmY is also a technique and exploits the statistical distribution of exponents in tensors. It can be used to quantize weights, static and dynamic activations, gradients, master weights and optimizer state. It can reduce memory (CPU DRAM and accelerator HBM), network and disk storage and transfers. It can increase multi tenancy and accelerate compute. eXmY has been deployed in production for almost 2 years.
FPGA-based hardware accelerators are becoming increasingly popular due to their versatility, customizability, energy efficiency, constant latency, and scalability. FPGAs can be tailored to specific algorithms, enabling efficient hardware implementations that effectively leverage algorithm parallelism. This can lead to significant performance improvements over CPUs and GPUs, particularly for highly parallel applications. For example, a recent study found that Stratix 10 FPGAs can achieve up to 90\% of the performance of a TitanX Pascal GPU while consuming less than 50\% of the power. This makes FPGAs an attractive choice for accelerating machine learning (ML) workloads. However, our research finds privacy and security vulnerabilities in existing Xilinx FPGA-based hardware acceleration solutions. These vulnerabilities arise from the lack of memory initialization and insufficient process isolation, which creates potential avenues for unauthorized access to private data used by processes. To illustrate this issue, we conducted experiments using a Xilinx ZCU104 board running the PetaLinux tool from Xilinx. We found that PetaLinux does not effectively clear memory locations associated with a terminated process, leaving them vulnerable to memory scraping attack (MSA). This paper makes two main contributions. The first contribution is an attack methodology of using the Xilinx debugger from a different user space. We find that we are able to access process IDs, virtual address spaces, and pagemaps of one user from a different user space because of lack of adequate process isolation. The second contribution is a methodology for characterizing terminated processes and accessing their private data. We illustrate this on Xilinx ML application library.
Automated driving systems are an integral part of the automotive industry. Tools such as Robot Operating System and simulators support their development. However, in the end, the developers must test their algorithms on a real vehicle. To better observe the difference between reality and simulation--the reality gap--digital twin technology offers real-time communication between the real vehicle and its model. We present low fidelity digital twin generator and describe situations where automatic generation is preferable to high fidelity simulation. We validated our approach of generating a virtual environment with a vehicle model by replaying the data recorded from the real vehicle.
While large language models (LLMs) have demonstrated remarkable capabilities across a range of downstream tasks, a significant concern revolves around their propensity to exhibit hallucinations: LLMs occasionally generate content that diverges from the user input, contradicts previously generated context, or misaligns with established world knowledge. This phenomenon poses a substantial challenge to the reliability of LLMs in real-world scenarios. In this paper, we survey recent efforts on the detection, explanation, and mitigation of hallucination, with an emphasis on the unique challenges posed by LLMs. We present taxonomies of the LLM hallucination phenomena and evaluation benchmarks, analyze existing approaches aiming at mitigating LLM hallucination, and discuss potential directions for future research.
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191