Dialog policies, which determine a system's action based on the current state at each dialog turn, are crucial to the success of the dialog. In recent years, reinforcement learning (RL) has emerged as a promising option for dialog policy learning (DPL). In RL-based DPL, dialog policies are updated according to rewards. The manual construction of fine-grained rewards, such as state-action-based ones, to effectively guide the dialog policy is challenging in multi-domain task-oriented dialog scenarios with numerous state-action pair combinations. One way to estimate rewards from collected data is to train the reward estimator and dialog policy simultaneously using adversarial learning (AL). Although this method has demonstrated superior performance experimentally, it is fraught with the inherent problems of AL, such as mode collapse. This paper first identifies the role of AL in DPL through detailed analyses of the objective functions of dialog policy and reward estimator. Next, based on these analyses, we propose a method that eliminates AL from reward estimation and DPL while retaining its advantages. We evaluate our method using MultiWOZ, a multi-domain task-oriented dialog corpus.
相關內容
We describe a path to humanity safely thriving with powerful Artificial General Intelligences (AGIs) by building them to provably satisfy human-specified requirements. We argue that this will soon be technically feasible using advanced AI for formal verification and mechanistic interpretability. We further argue that it is the only path which guarantees safe controlled AGI. We end with a list of challenge problems whose solution would contribute to this positive outcome and invite readers to join in this work.
Stein's method for Gaussian process approximation can be used to bound the differences between the expectations of smooth functionals $h$ of a c\`adl\`ag random process $X$ of interest and the expectations of the same functionals of a well understood target random process $Z$ with continuous paths. Unfortunately, the class of smooth functionals for which this is easily possible is very restricted. Here, we prove an infinite dimensional Gaussian smoothing inequality, which enables the class of functionals to be greatly expanded -- examples are Lipschitz functionals with respect to the uniform metric, and indicators of arbitrary events -- in exchange for a loss of precision in the bounds. Our inequalities are expressed in terms of the smooth test function bound, an expectation of a functional of $X$ that is closely related to classical tightness criteria, a similar expectation for $Z$, and, for the indicator of a set $K$, the probability $\mathbb{P}(Z \in K^\theta \setminus K^{-\theta})$ that the target process is close to the boundary of $K$.
In swarm robotics, agents interact through local roles to solve complex tasks beyond an individual's ability. Even though swarms are capable of carrying out some operations without the need for human intervention, many safety-critical applications still call for human operators to control and monitor the swarm. There are novel challenges to effective Human-Swarm Interaction (HSI) that are only beginning to be addressed. Explainability is one factor that can facilitate effective and trustworthy HSI and improve the overall performance of Human-Swarm team. Explainability was studied across various Human-AI domains, such as Human-Robot Interaction and Human-Centered ML. However, it is still ambiguous whether explanations studied in Human-AI literature would be beneficial in Human-Swarm research and development. Furthermore, the literature lacks foundational research on the prerequisites for explainability requirements in swarm robotics, i.e., what kind of questions an explainable swarm is expected to answer, and what types of explanations a swarm is expected to generate. By surveying 26 swarm experts, we seek to answer these questions and identify challenges experts faced to generate explanations in Human-Swarm environments. Our work contributes insights into defining a new area of research of eXplainable Swarm (xSwarm) which looks at how explainability can be implemented and developed in swarm systems. This paper opens the discussion on xSwarm and paves the way for more research in the field.
Electrical circuits are present in a variety of technologies, making their design an important part of computer aided engineering. The growing number of tunable parameters that affect the final design leads to a need for new approaches of quantifying their impact. Machine learning may play a key role in this regard, however current approaches often make suboptimal use of existing knowledge about the system at hand. In terms of circuits, their description via modified nodal analysis is well-understood. This particular formulation leads to systems of differential-algebraic equations (DAEs) which bring with them a number of peculiarities, e.g. hidden constraints that the solution needs to fulfill. We aim to use the recently introduced dissection concept for DAEs that can decouple a given system into ordinary differential equations, only depending on differential variables, and purely algebraic equations that describe the relations between differential and algebraic variables. The idea then is to only learn the differential variables and reconstruct the algebraic ones using the relations from the decoupling. This approach guarantees that the algebraic constraints are fulfilled up to the accuracy of the nonlinear system solver, which represents the main benefit highlighted in this article.
Understanding the mechanism of how convolutional neural networks learn features from image data is a fundamental problem in machine learning and computer vision. In this work, we identify such a mechanism. We posit the Convolutional Neural Feature Ansatz, which states that covariances of filters in any convolutional layer are proportional to the average gradient outer product (AGOP) taken with respect to patches of the input to that layer. We present extensive empirical evidence for our ansatz, including identifying high correlation between covariances of filters and patch-based AGOPs for convolutional layers in standard neural architectures, such as AlexNet, VGG, and ResNets pre-trained on ImageNet. We also provide supporting theoretical evidence. We then demonstrate the generality of our result by using the patch-based AGOP to enable deep feature learning in convolutional kernel machines. We refer to the resulting algorithm as (Deep) ConvRFM and show that our algorithm recovers similar features to deep convolutional networks including the notable emergence of edge detectors. Moreover, we find that Deep ConvRFM overcomes previously identified limitations of convolutional kernels, such as their inability to adapt to local signals in images and, as a result, leads to sizable performance improvement over fixed convolutional kernels.
We provide a new sequent calculus that enjoys syntactic cut-elimination and strongly terminating backward proof search for the intuitionistic Strong L\"ob logic $\sf{iSL}$, an intuitionistic modal logic with a provability interpretation. A novel measure on sequents is used to prove both the termination of the naive backward proof search strategy, and the admissibility of cut in a syntactic and direct way, leading to a straightforward cut-elimination procedure. All proofs have been formalised in the interactive theorem prover Coq.
We present Carnot, a leader-based Byzantine Fault Tolerant (BFT) consensus protocol that is responsive and operates under the partially synchronous model. Responsive BFT consensus protocols exhibit wire-speed operation and deliver instantaneous finality, thereby addressing a fundamental need in distributed systems. A key challenge in scaling these protocols has been the computational complexity associated with authenticator verification. We demonstrate that Carnot effectively addresses this bottleneck by adeptly streamlining the verification and aggregation of $O(log(N))$ authenticators per node. This notable advancement marks a substantial improvement over the prevailing $O(N)$ state-of-the-art approaches. Leveraging this inherent property, Carnot demonstrates its capacity to seamlessly scale to networks comprising tens to hundreds of thousands of nodes. We envision Carnot as a critical stride towards bridging the gap between classical BFT consensus mechanisms and blockchain technology.
Graph Neural Networks (GNNs) are becoming increasingly popular due to their superior performance in critical graph-related tasks. While quantization is widely used to accelerate GNN computation, quantized training faces unprecedented challenges. Current quantized GNN training systems often have longer training times than their full-precision counterparts for two reasons: (i) addressing the accuracy challenge leads to excessive overhead, and (ii) the optimization potential exposed by quantization is not adequately leveraged. This paper introduces Tango which re-thinks quantization challenges and opportunities for graph neural network training on GPUs with three contributions: Firstly, we introduce efficient rules to maintain accuracy during quantized GNN training. Secondly, we design and implement quantization-aware primitives and inter-primitive optimizations that can speed up GNN training. Finally, we integrate Tango with the popular Deep Graph Library (DGL) system and demonstrate its superior performance over state-of-the-art approaches on various GNN models and datasets.
Artificial neural networks thrive in solving the classification problem for a particular rigid task, acquiring knowledge through generalized learning behaviour from a distinct training phase. The resulting network resembles a static entity of knowledge, with endeavours to extend this knowledge without targeting the original task resulting in a catastrophic forgetting. Continual learning shifts this paradigm towards networks that can continually accumulate knowledge over different tasks without the need to retrain from scratch. We focus on task incremental classification, where tasks arrive sequentially and are delineated by clear boundaries. Our main contributions concern 1) a taxonomy and extensive overview of the state-of-the-art, 2) a novel framework to continually determine the stability-plasticity trade-off of the continual learner, 3) a comprehensive experimental comparison of 11 state-of-the-art continual learning methods and 4 baselines. We empirically scrutinize method strengths and weaknesses on three benchmarks, considering Tiny Imagenet and large-scale unbalanced iNaturalist and a sequence of recognition datasets. We study the influence of model capacity, weight decay and dropout regularization, and the order in which the tasks are presented, and qualitatively compare methods in terms of required memory, computation time, and storage.
Meta-learning, or learning to learn, has gained renewed interest in recent years within the artificial intelligence community. However, meta-learning is incredibly prevalent within nature, has deep roots in cognitive science and psychology, and is currently studied in various forms within neuroscience. The aim of this review is to recast previous lines of research in the study of biological intelligence within the lens of meta-learning, placing these works into a common framework. More recent points of interaction between AI and neuroscience will be discussed, as well as interesting new directions that arise under this perspective.
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191