Self-consistency-based approaches, which involve repeatedly sampling multiple outputs and selecting the most consistent one as the final response, prove to be remarkably effective in improving the factual accuracy of large language models. Nonetheless, existing methods usually have strict constraints on the task format, largely limiting their applicability. In this paper, we present Integrative Decoding (ID), to unlock the potential of self-consistency in open-ended generation tasks. ID operates by constructing a set of inputs, each prepended with a previously sampled response, and then processes them concurrently, with the next token being selected by aggregating of all their corresponding predictions at each decoding step. In essence, this simple approach implicitly incorporates self-consistency in the decoding objective. Extensive evaluation shows that ID consistently enhances factuality over a wide range of language models, with substantial improvements on the TruthfulQA (+11.2%), Biographies (+15.4%) and LongFact (+8.5%) benchmarks. The performance gains amplify progressively as the number of sampled responses increases, indicating the potential of ID to scale up with repeated sampling.

In recent years, there has been a growing interest in exploring dialogues with more complex goals, such as negotiation, persuasion, and emotional support, which go beyond traditional service-focused dialogue systems. Apart from the requirement for much more sophisticated strategic reasoning and communication skills, a significant challenge of these tasks lies in the difficulty of objectively measuring the achievement of their goals in a quantifiable way, making it difficult for existing research to directly optimize the dialogue procedure towards them. In our work, we emphasize the multifaceted nature of complex dialogue goals and argue that it is more feasible to accomplish them by comprehensively considering and jointly promoting their different aspects. To this end, we propose a novel dialogue framework, Cooper, which coordinates multiple specialized agents, each dedicated to a specific dialogue goal aspect separately, to approach the complex objective. Through this divide-and-conquer manner, we make complex dialogue goals more approachable and elicit greater intelligence via the collaboration of individual agents. Experiments on persuasion and emotional support dialogues demonstrate the superiority of our method over a set of competitive baselines. Our codes are available at https://github.com/YiCheng98/Cooper.

When seismic waves propagate through hydrocarbon reservoirs, there is a significant attenuation of high-frequency energy. The quality factor Q is usually treated as a sensitive parameter indicating oil and gas reservoirs, but extracting seismic attenuation parameters using the conventional time-frequency domain methods often yields unsatisfactory results. In order to improve the accuracy of Q estimation, we propose a new method for calculating seismic parameters. We first introduce a Gaussian window into the general linear chirplet transform (GLCT) based on dominant frequency weight, which is now used in the W-transform (WT). This modification enables the improved GLCT (IGLCT) to adjust its time-frequency resolution according to the changes at the seismic dominant frequency, thereby enhancing the time resolution of conventional GLCT. Subsequently, based on the characteristics of high-frequency energy attenuation anomalies in seismic records and in combination with the Teager-Kaiser energy operator (TKEO), then the effective estimation of seismic attenuation parameters is achieved. This method can obtain a higher resolution time-frequency spectrum compared to conventional time-frequency analysis (TFA) methods. In addition, the TKEO used in this technique has significant instantaneous features, which can be used to accurately estimate the energy of monofrequency signals. Finally, seismic attenuation parameters are calculated based on the quantified definition of the quality factor Q and the relative energy attenuation within the given time window. Theoretical model testing and actual data processing indicate that the proposed method for estimating seismic attenuation parameters has the capability for identifying the thin layers and can effectively predict the hydrocarbon reservoirs.

Multihorizon tracking deep learning methods have higher efficiency than single-horizon tracking deep learning methods, but they still have a great challenge to adapt to complex seismic data, resulting in undesired horizon tracking results. To achieve high-precision horizon tracking results, we propose a boundary-constrained multitask multihorizon tracking (BMTM) based on semantic segmentation, instance segmentation, and multitask learning (ML). The core idea of BMTM is to transform multihorizon tracking into strata recognition, using where the strata boundaries serve as the tracking results. BMTM comprises three components: strata label automatic generation, strata boundary-constrained multihorizon tracking network (BMTN), and strata boundary-constrained loss function. Strata label automatic generation automatically generates strata labels based on the input horizon labels. BMTN consists of a shared layer, an auxiliary task, and a main task. The auxiliary task takes the input horizon labels as its labels and employs semantic segmentation to directly output multihorizon tracking results. Main task takes the generated strata labels as its labels and employs instance segmentation with the outputs of auxiliary task to achieve high-precision strata (or horizon) tracking results. The strata boundary-constrained loss function aims to pay more attention to the strata boundary and ultimately improve the horizon tracking precision. One public 3-D synthetic seismic dataset study demonstrated the performance of BMTM, and one field 3-D seismic dataset application demonstrated that BMTM can be used for high-precision multihorizon tracking.

In the oil and gas exploration, delineating the tectonic structure of geological targets and predicting physical properties are paramount. As we all know, seismic velocity modeling is crucial for characterizing reservoirs. Enhancing seismic inversion accuracy to obtain quantitative parameters is vital in oil reservoir exploration. Current seismic waveform inversion utilizes the amplitude and timing of seismic records to reconstruct the velocity model, thereby refining the model for seismic migration. This article addresses the convergence and computational challenges in conventional inversion by proposing an adaptive iteration method for step updates. This method outperforms inexact linear searches and empirical formulas by updating the model effectively without additional seismic simulations. In the past, gradient optimization typically requires three seismic modeling in a single iteration, whereas adaptive iteration only requires two, offering practical cost savings. To address stability issues in complex scenarios, we combine parabolic search with adaptive iteration, which is the modified optimization technique, to enhance the convergence rate of objective function. Furthermore, by employing gradient preconditioning based on illumination and multigrid strategy, we achieve full strata velocity modeling from shallow to deep layers. Numerical tests on the theoretical anomaly model and BP model validate the effectiveness and precision of the adaptive iteration, significantly reducing the computational load of seismic inversion and accurately estimating deep strata information.

Providing Emotional Support (ES) to soothe people in emotional distress is an essential capability in social interactions.Most existing research on building ES conversation systems only considers single-turn interactions with users, which is over-simplified.In comparison, multi-turn ES conversation systems can provide ES more effectively, but face several new technical challenges, including: i) how to conduct support strategy planning that could lead to the best supporting effects; ii) how to dynamically model the user's state.In this paper, we propose a novel system named MultiESC to address these issues.For strategy planning, drawing inspiration from the A* search algorithm, we propose lookahead heuristics to estimate the future user feedback after using particular strategies, which helps to select strategies that can lead to the best long-term effects.For user state modeling, MultiESC focuses on capturing users' subtle emotional expressions and understanding their emotion causes.Extensive experiments show that MultiESC significantly outperforms competitive baselines in both strategy planning and dialogue generation.Our codes are available at https: //github.com/lwgkzl/MultiESC.

With the rise of telemedicine, the task of developing Dialogue Systems for Medical Diagnosis (DSMD) has received much attention in recent years. Different from early researches that needed to rely on extra human resources and expertise to build the system, recent researches focused on how to build DSMD in a data-driven manner. However, the previous data-driven DSMD methods largely overlooked the system interpretability, which is critical for a medical application, and they also suffered from the data sparsity issue at the same time. In this paper, we explore how to bring interpretability to data-driven DSMD. Specifically, we propose a more interpretable decision process to implement the dialogue manager of DSMD by reasonably mimicking real doctors' inquiry logics, and we devise a model with highly transparent components to conduct the inference. Moreover, we collect a new DSMD dataset, which has a much larger scale, more diverse patterns, and is of higher quality than the existing ones. The experiments show that our method obtains 7.7%, 10.0%, 3.0% absolute improvement in diagnosis accuracy respectively on three datasets, demonstrating the effectiveness of its rational decision process and model design. Our codes and the GMD-12 dataset are available at https://github.com/lwgkzl/BR-Agent.