Current literature, aiming to surpass the "Chain-of-Thought" approach, often resorts to an external modus operandi involving halting, modifying, and then resuming the generation process to boost Large Language Models' (LLMs) reasoning capacities. This mode escalates the number of query requests, leading to increased costs, memory, and computational overheads. Addressing this, we propose the Algorithm of Thoughts -- a novel strategy that propels LLMs through algorithmic reasoning pathways, pioneering a new mode of in-context learning. By employing algorithmic examples, we exploit the innate recurrence dynamics of LLMs, expanding their idea exploration with merely one or a few queries. Our technique outperforms earlier single-query methods and stands on par with a recent multi-query strategy that employs an extensive tree search algorithm. Intriguingly, our results suggest that instructing an LLM using an algorithm can lead to performance surpassing that of the algorithm itself, hinting at LLM's inherent ability to weave its intuition into optimized searches. We probe into the underpinnings of our method's efficacy and its nuances in application.
相關內容
Fine-tuning diffusion models through personalized datasets is an acknowledged method for improving generation quality across downstream tasks, which, however, often inadvertently generates unintended concepts such as watermarks and QR codes, attributed to the limitations in image sources and collecting methods within specific downstream tasks. Existing solutions suffer from eliminating these unintentionally learned implicit concepts, primarily due to the dependency on the model's ability to recognize concepts that it actually cannot discern. In this work, we introduce \methodname, a novel approach that successfully removes the implicit concepts with either an additional accessible classifier or detector model to encode geometric information of these concepts into text domain. Moreover, we propose \textit{Implicit Concept}, a novel image-text dataset imbued with three implicit concepts (\ie, watermarks, QR codes, and text) for training and evaluation. Experimental results demonstrate that \methodname not only identifies but also proficiently eradicates implicit concepts, revealing a significant improvement over the existing methods. The integration of geometric information marks a substantial progression in the precise removal of implicit concepts in diffusion models.
There is a fierce competition between two-sided mobility platforms (e.g., Uber and Lyft) fueled by massive subsidies, yet the underlying dynamics and interactions between the competing plat-forms are largely unknown. These platforms rely on the cross-side network effects to grow, they need to attract agents from both sides to kick-off: travellers are needed for drivers and drivers are needed for travellers. We use our coevolutionary model featured by the S-shaped learning curves to simulate the day-to-day dynamics of the ride-sourcing market at the microscopic level. We run three scenarios to illustrate the possible equilibria in the market. Our results underline how the correlation inside the ride-sourcing nest of the agents choice set significantly affects the plat-forms' market shares. While late entry to the market decreases the chance of platform success and possibly results in "winner-takes-all", heavy subsidies can keep the new platform in competition giving rise to "market sharing" regime.
We consider nonparametric estimation in Wicksell's problem and we show that the isotonized version of the plug-in estimator is asymptotically efficient. The asymptotic variance will depend on the local smoothness at the estimation point and at $0$ of the unknown distribution function $F$ of the ball-squared radii. We prove a corresponding local asymptotic minimax lower bound, as well under some local smoothness condition. This solves in an adaptive way the nonparametric estimation problem.
Blockchain provides decentralization and trustlessness features for the Industrial Internet of Things (IIoT), which expands the application scenarios of IIoT. To address the problem that the blockchain cannot actively obtain off-chain data, the blockchain oracle is proposed as a bridge between the blockchain and external data. However, the existing oracle schemes are difficult to solve the problem of low quality of service caused by frequent data changes and heterogeneous devices in IIoT, and the current oracle node selection schemes are difficult to balance security and quality of service. To tackle these problems, this paper proposes a secure and reliable oracle scheme that can obtain high-quality off-chain data. Specifically, we first design an oracle node selection algorithm based on Verifiable Random Function (VRF) and reputation mechanism to securely select high-quality nodes. Second, we propose a data filtering algorithm based on a sliding window to further improve the consistency of the collected data. We verify the security of the proposed scheme through security analysis. The experimental results show that the proposed scheme can effectively improve the service quality of the oracle.
Objective: For lower arm amputees, robotic prosthetic hands promise to regain the capability to perform daily living activities. Current control methods based on physiological signals such as electromyography (EMG) are prone to yielding poor inference outcomes due to motion artifacts, muscle fatigue, and many more. Vision sensors are a major source of information about the environment state and can play a vital role in inferring feasible and intended gestures. However, visual evidence is also susceptible to its own artifacts, most often due to object occlusion, lighting changes, etc. Multimodal evidence fusion using physiological and vision sensor measurements is a natural approach due to the complementary strengths of these modalities. Methods: In this paper, we present a Bayesian evidence fusion framework for grasp intent inference using eye-view video, eye-gaze, and EMG from the forearm processed by neural network models. We analyze individual and fused performance as a function of time as the hand approaches the object to grasp it. For this purpose, we have also developed novel data processing and augmentation techniques to train neural network components. Results: Our results indicate that, on average, fusion improves the instantaneous upcoming grasp type classification accuracy while in the reaching phase by 13.66% and 14.8%, relative to EMG and visual evidence individually, resulting in an overall fusion accuracy of 95.3%. Conclusion: Our experimental data analyses demonstrate that EMG and visual evidence show complementary strengths, and as a consequence, fusion of multimodal evidence can outperform each individual evidence modality at any given time.
Large Language Models (LLMs) have the ability to solve a variety of tasks, such as text summarization and mathematical questions, just out of the box, but they are often trained with a single task in mind. Due to high computational costs, the current trend is to use prompt instruction tuning to better adjust monolithic, pretrained LLMs for new -- but often individual -- downstream tasks. Thus, how one would expand prompt tuning to handle -- concomitantly -- heterogeneous tasks and data distributions is a widely open question. To address this gap, we suggest the use of \emph{Mixture of Prompts}, or MoPs, associated with smart gating functionality: the latter -- whose design is one of the contributions of this paper -- can identify relevant skills embedded in different groups of prompts and dynamically assign combined experts (i.e., collection of prompts), based on the target task. Additionally, MoPs are empirically agnostic to any model compression technique applied -- for efficiency reasons -- as well as instruction data source and task composition. In practice, MoPs can simultaneously mitigate prompt training "interference" in multi-task, multi-source scenarios (e.g., task and data heterogeneity across sources), as well as possible implications from model approximations. As a highlight, MoPs manage to decrease final perplexity from $\sim20\%$ up to $\sim70\%$, as compared to baselines, in the federated scenario, and from $\sim 3\%$ up to $\sim30\%$ in the centralized scenario.
Connected Medical Devices (CMDs) have a large impact on patients as they allow them to lead a more normal life. Any malfunction could not only remove the health benefits the CMDs provide, they could also cause further harm to the patient. Due to this, there are many safety regulations which must be adhered to prior to a CMD entering the market. However, while many detailed safety regulations exist, there are a fundamental lack of cybersecurity frameworks applicable to CMDs. While there are recent regulations which aim to enforce cybersecurity practices, they are vague and do not contain the concrete steps necessary to implement cybersecurity. This paper aims to fill that gap by describing a framework, CyMed, to be used by vendors and ens-users, which contains concrete measures to improve the resilience of CMDs against cyber attack. The CyMed framework is subsequently evaluated based on practical tests as well as expert interviews.
Differentiable logics (DL) have recently been proposed as a method of training neural networks to satisfy logical specifications. A DL consists of a syntax in which specifications are stated and an interpretation function that translates expressions in the syntax into loss functions. These loss functions can then be used during training with standard gradient descent algorithms. The variety of existing DLs and the differing levels of formality with which they are treated makes a systematic comparative study of their properties and implementations difficult. This paper remedies this problem by suggesting a meta-language for defining DLs that we call the Logic of Differentiable Logics, or LDL. Syntactically, it generalises the syntax of existing DLs to FOL, and for the first time introduces the formalism for reasoning about vectors and learners. Semantically, it introduces a general interpretation function that can be instantiated to define loss functions arising from different existing DLs. We use LDL to establish several theoretical properties of existing DLs, and to conduct their empirical study in neural network verification.
Increasing the context length of large language models (LLMs) unlocks fundamentally new capabilities, but also significantly increases the memory footprints of training. Previous model-parallel systems such as Megatron-LM partition and compute different attention heads in parallel, resulting in large communication volumes, so they cannot scale beyond the number of attention heads, thereby hindering its adoption. In this paper, we introduce a new approach, LightSeq, for long-context LLMs training. LightSeq has many notable advantages. First, LightSeq partitions over the sequence dimension, hence is agnostic to model architectures and readily applicable for models with varying numbers of attention heads, such as Multi-Head, Multi-Query and Grouped-Query attention. Second, LightSeq not only requires up to 4.7x less communication than Megatron-LM on popular LLMs but also overlaps the communication with computation. To further reduce the training time, LightSeq features a novel gradient checkpointing scheme to bypass an forward computation for memory-efficient attention. We evaluate LightSeq on Llama-7B and its variants with sequence lengths from 32K to 512K. Through comprehensive experiments on single and cross-node training, we show that LightSeq achieves up to 1.24-2.01x end-to-end speedup, and a 2-8x longer sequence length on models with fewer heads, compared to Megatron-LM. Codes will be available at //github.com/RulinShao/LightSeq.
Graph Neural Networks (GNNs) have been studied from the lens of expressive power and generalization. However, their optimization properties are less well understood. We take the first step towards analyzing GNN training by studying the gradient dynamics of GNNs. First, we analyze linearized GNNs and prove that despite the non-convexity of training, convergence to a global minimum at a linear rate is guaranteed under mild assumptions that we validate on real-world graphs. Second, we study what may affect the GNNs' training speed. Our results show that the training of GNNs is implicitly accelerated by skip connections, more depth, and/or a good label distribution. Empirical results confirm that our theoretical results for linearized GNNs align with the training behavior of nonlinear GNNs. Our results provide the first theoretical support for the success of GNNs with skip connections in terms of optimization, and suggest that deep GNNs with skip connections would be promising in practice.
小貼士
登錄享
相關主題
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191