Search

Large language models has catalyzed the development of personalized dialogue systems, numerous role-playing conversational agents have emerged. While previous research predominantly focused on enhancing the model's capability to follow instructions by designing character profiles, neglecting the psychological factors that drive human conversations. In this paper, we propose Orca, a framework for data processing and training LLMs of custom characters by integrating personality traits. Orca comprises four stages: (1) Personality traits inferring, leverage LLMs to infer user's BigFive personality trait reports and scores. (2) Data Augment, simulate user's profile, background story, and psychological activities. (3) Dataset construction, personality-conditioned instruction prompting (PCIP) to stimulate LLMs. (4) Modeling and Training, personality-conditioned instruction tuning (PTIT and PSIT), using the generated data to enhance existing open-source LLMs. We introduce OrcaBench, the first benchmark for evaluating the quality of content generated by LLMs on social platforms across multiple scales. Our experiments demonstrate that our proposed model achieves superior performance on this benchmark, demonstrating its excellence and effectiveness in perceiving personality traits that significantly improve role-playing abilities. Our Code is available at https://github.com/Aipura/Orca.

This paper presents a method to evaluate the alignment between the decision-making logic of Large Language Models (LLMs) and human cognition in a case study on legal LLMs. Unlike traditional evaluations on language generation results, we propose to evaluate the correctness of the detailed decision-making logic of an LLM behind its seemingly correct outputs, which represents the core challenge for an LLM to earn human trust. To this end, we quantify the interactions encoded by the LLM as primitive decision-making logic, because recent theoretical achievements have proven several mathematical guarantees of the faithfulness of the interaction-based explanation. We design a set of metrics to evaluate the detailed decision-making logic of LLMs. Experiments show that even when the language generation results appear correct, a significant portion of the internal inference logic contains notable issues.

In Video-based Facial Expression Recognition (V-FER), models are typically trained on closed-set datasets with a fixed number of known classes. However, these models struggle with unknown classes common in real-world scenarios. In this paper, we introduce a challenging Open-set Video-based Facial Expression Recognition (OV-FER) task, aiming to identify both known and new, unseen facial expressions. While existing approaches use large-scale vision-language models like CLIP to identify unseen classes, we argue that these methods may not adequately capture the subtle human expressions needed for OV-FER. To address this limitation, we propose a novel Human Expression-Sensitive Prompting (HESP) mechanism to significantly enhance CLIP's ability to model video-based facial expression details effectively. Our proposed HESP comprises three components: 1) a textual prompting module with learnable prompts to enhance CLIP's textual representation of both known and unknown emotions, 2) a visual prompting module that encodes temporal emotional information from video frames using expression-sensitive attention, equipping CLIP with a new visual modeling ability to extract emotion-rich information, and 3) an open-set multi-task learning scheme that promotes interaction between the textual and visual modules, improving the understanding of novel human emotions in video sequences. Extensive experiments conducted on four OV-FER task settings demonstrate that HESP can significantly boost CLIP's performance (a relative improvement of 17.93% on AUROC and 106.18% on OSCR) and outperform other state-of-the-art open-set video understanding methods by a large margin. Code is available at https://github.com/cosinehuang/HESP., Comment: Accepted by ACM MM2024

Recent advancements in multi-modal artificial intelligence (AI) have revolutionized the fields of stock market forecasting and heart rate monitoring. Utilizing diverse data sources can substantially improve prediction accuracy. Nonetheless, additional data may not always align with the original dataset. Interpolation methods are commonly utilized for handling missing values in modal data, though they may exhibit limitations in the context of sparse information. Addressing this challenge, we propose a Modality Completion Deep Belief Network-Based Model (MC-DBN). This approach utilizes implicit features of complete data to compensate for gaps between itself and additional incomplete data. It ensures that the enhanced multi-modal data closely aligns with the dynamic nature of the real world to enhance the effectiveness of the model. We conduct evaluations of the MC-DBN model in two datasets from the stock market forecasting and heart rate monitoring domains. Comprehensive experiments showcase the model's capacity to bridge the semantic divide present in multi-modal data, subsequently enhancing its performance. The source code is available at: https://github.com/logan-0623/DBN-generate

The development of large language models (LLMs) has been catalyzed by advancements in pre-training techniques. These models have demonstrated robust reasoning capabilities through manually designed prompts. In this work, we evaluate the conversational reasoning capabilities of the current state-of-the-art LLM (GPT-4) on knowledge graphs (KGs). However, the performance of LLMs is constrained due to a lack of KG environment awareness and the difficulties in developing effective optimization mechanisms for intermediary reasoning stages. We further introduce LLM-ARK, a LLM grounded KG reasoning agent designed to deliver precise and adaptable predictions on KG paths. LLM-ARK leverages Full Textual Environment (FTE) prompt to assimilate state information within each reasoning step. We reframe the challenge of multi-hop reasoning on the KG as a sequential decision-making task. Utilizing the Proximal Policy Optimization (PPO) online policy gradient reinforcement learning algorithm, our model is optimized to learn from rich reward signals. Additionally, we conduct an evaluation of our model and GPT-4 on the OpenDialKG dataset. The experimental results reveal that LLaMA-2-7B-ARK outperforms the current state-of-the-art model by 5.28 percentage points, with a performance rate of 36.39% on the target@1 evaluation metric. Meanwhile, GPT-4 scored 14.91%, further demonstrating the effectiveness of our method. Our code is available on GitHub (https://github.com/Aipura/LLM-ARK) for further access.

In this paper, we propose a novel, computationally efficient reduced order method to solve linear parabolic inverse source problems. Our approach provides accurate numerical solutions without relying on specific training data. The forward solution is constructed using a Krylov sequence, while the source term is recovered via the conjugate gradient (CG) method. Under a weak regularity assumption on the solution of the parabolic partial differential equations (PDEs), we establish convergence of the forward solution and provide a rigorous error estimate for our method. Numerical results demonstrate that our approach offers substantial computational savings compared to the traditional finite element method (FEM) and retains equivalent accuracy., Comment: This is a placeholder. Unfinished Section 4 and Section 6

The field of image blending has gained popularity in recent years for its ability to create visually stunning content. However, the current image blending algorithm has the following problems: 1) The manual creation of the image blending mask requires a lot of manpower and material resources; 2) The image blending algorithm cannot effectively solve the problems of brightness distortion and low resolution. To this end, we propose a new image blending method: it combines semantic object detection and segmentation with corresponding mask generation to automatically blend images, while a two-stage iterative algorithm based on our proposed new saturation loss and PAN algorithm to fix brightness distortion and low resolution issues. Results on publicly available datasets show that our method outperforms many classic image blending algorithms on various performance metrics such as PSNR and SSIM., Comment: 15 pages, 9 figure

Hall effect measurements in doped polymer semiconductors are widely reported, but are difficult to interpret due to screening of Hall voltages by carriers undergoing incoherent transport. Here, we propose a refined analysis for such Hall measurements, based on measuring the Hall coefficient as a function of temperature, and modelling carriers as existing in a regime of variable "deflectability" (i.e. how strongly they "feel" the magnetic part of the Lorentz force). By linearly interpolating each carrier between the extremes of no deflection and full deflection, we demonstrate that it is possible to extract the (time-averaged) concentration of deflectable charge carriers, $\left$, the average, temperature-dependent mobility of those carriers, $\left<\mu_d\right>(T)$, as well as the ratio of conductivity that comes from such deflectable transport, $d(T)$. Our method was enabled by the construction of an improved AC Hall measurement system, as well as an improved data extraction method. We measured Hall bar devices of ion-exchange doped films of PBTTT-C$_{14}$ from 10--300 K. Our analysis provides evidence for the proportion of conductivity arising from deflectable transport, $d(T)$, increasing with doping level, ranging between 15.4% and 16.4% at room temperature. When compared to total charge-carrier-density estimates from independent methods, the values of $\left$ extracted suggest that carriers spend $\sim$37% of their time of flight being deflectable in the most highly doped of the devices measured here. The extracted values of $d(T)$ being less than half this value thus suggest that the limiting factor for conductivity in such highly doped devices is carrier mobility, rather than concentration., Comment: Accepted for publication in APS Physical Review Materials

Crash data of autonomous vehicles (AV) or vehicles equipped with advanced driver assistance systems (ADAS) are the key information to understand the crash nature and to enhance the automation systems. However, most of the existing crash data sources are either limited by the sample size or suffer from missing or unverified data. To contribute to the AV safety research community, we introduce AVOID: an open AV crash dataset. Three types of vehicles are considered: Advanced Driving System (ADS) vehicles, Advanced Driver Assistance Systems (ADAS) vehicles, and low-speed autonomous shuttles. The crash data are collected from the National Highway Traffic Safety Administration (NHTSA), California Department of Motor Vehicles (CA DMV) and incident news worldwide, and the data are manually verified and summarized in ready-to-use format. In addition, land use, weather, and geometry information are also provided. The dataset is expected to accelerate the research on AV crash analysis and potential risk identification by providing the research community with data of rich samples, diverse data sources, clear data structure, and high data quality.

Application of machine learning for stock prediction is attracting a lot of attention in recent years. A large amount of research has been conducted in this area and multiple existing results have shown that machine learning methods could be successfully used toward stock predicting using stocks historical data. Most of these existing approaches have focused on short term prediction using stocks historical price and technical indicators. In this paper, we prepared 22 years worth of stock quarterly financial data and investigated three machine learning algorithms: Feed-forward Neural Network (FNN), Random Forest (RF) and Adaptive Neural Fuzzy Inference System (ANFIS) for stock prediction based on fundamental analysis. In addition, we applied RF based feature selection and bootstrap aggregation in order to improve model performance and aggregate predictions from different models. Our results show that RF model achieves the best prediction results, and feature selection is able to improve test performance of FNN and ANFIS. Moreover, the aggregated model outperforms all baseline models as well as the benchmark DJIA index by an acceptable margin for the test period. Our findings demonstrate that machine learning models could be used to aid fundamental analysts with decision-making regarding stock investment., Comment: 10 pages. IEEE Symposium Series on Computational Intelligence, Orlando, Florida,USA, December 2021

Doped organic semiconductors are critical to emerging device applications, including thermoelectrics, bioelectronics, and neuromorphic computing devices. It is commonly assumed that low conductivities in these materials result primarily from charge trapping by the Coulomb potentials of the dopant counter-ions. Here, we present a combined experimental and theoretical study rebutting this belief. Using a newly developed doping technique, we find the conductivity of several classes of high-mobility conjugated polymers to be strongly correlated with paracrystalline disorder but poorly correlated with ionic size, suggesting that Coulomb traps do not limit transport. A general model for interacting electrons in highly doped polymers is proposed and carefully parameterized against atomistic calculations, enabling the calculation of electrical conductivity within the framework of transient localisation theory. Theoretical calculations are in excellent agreement with experimental data, providing insights into the disordered-limited nature of charge transport and suggesting new strategies to further improve conductivities.

Application of neural network architectures for financial prediction has been actively studied in recent years. This paper presents a comparative study that investigates and compares feed-forward neural network (FNN) and adaptive neural fuzzy inference system (ANFIS) on stock prediction using fundamental financial ratios. The study is designed to evaluate the performance of each architecture based on the relative return of the selected portfolios with respect to the benchmark stock index. The results show that both architectures possess the ability to separate winners and losers from a sample universe of stocks, and the selected portfolios outperform the benchmark. Our study argues that FNN shows superior performance over ANFIS., Comment: 4 pages