arXiv subjects

Shu-Ting Pi

Publications and source records attributed to Shu-Ting Pi.

13 recordsLinked to original sources

arXiv2026

Don't Shoot The Breeze: Topic Continuity Model Using Nonlinear Naive Bayes With Attention

Utilizing Large Language Models (LLM) as chatbots in diverse business scenarios often presents the challenge of maintaining topic continuity. Abrupt shifts in topics can lead to poor user experiences and inefficient utilization of computational resources. In this paper, we present a topic continuity model aimed at assessing whether a response aligns with the initial conversation topic. Our model is built upon the expansion of the corresponding natural language understanding (NLU) model into quantifiable terms using a Naive Bayes approach. Subsequently, we have introduced an attention mechanism and logarithmic nonlinearity to enhance its capability to capture topic continuity. This approach allows us to convert the NLU model into an interpretable analytical formula. In contrast to many NLU models constrained by token limits, our proposed model can seamlessly handle conversations of any length with linear time complexity. Furthermore, the attention mechanism significantly improves the model's ability to identify topic continuity in complex conversations. According to our experiments, our model consistently outperforms traditional methods, particularly in handling lengthy and intricate conversations. This unique capability offers us an opportunity to ensure the responsible and interpretable use of LLMs.

cs.CL↗

arXiv2025

REIC: RAG-Enhanced Intent Classification at Scale

Accurate intent classification is critical for efficient routing in customer service, ensuring customers are connected with the most suitable agents while reducing handling times and operational costs. However, as companies expand their product lines, intent classification faces scalability challenges due to the increasing number of intents and variations in taxonomy across different verticals. In this paper, we introduce REIC, a Retrieval-augmented generation Enhanced Intent Classification approach, which addresses these challenges effectively. REIC leverages retrieval-augmented generation (RAG) to dynamically incorporate relevant knowledge, enabling precise classification without the need for frequent retraining. Through extensive experiments on real-world datasets, we demonstrate that REIC outperforms traditional fine-tuning, zero-shot, and few-shot methods in large-scale customer service settings. Our results highlight its effectiveness in both in-domain and out-of-domain scenarios, demonstrating its potential for real-world deployment in adaptive and large-scale intent classification systems.

cs.CL↗

arXiv2024

Contact Complexity in Customer Service

Customers who reach out for customer service support may face a range of issues that vary in complexity. Routing high-complexity contacts to junior agents can lead to multiple transfers or repeated contacts, while directing low-complexity contacts to senior agents can strain their capacity to assist customers who need professional help. To tackle this, a machine learning model that accurately predicts the complexity of customer issues is highly desirable. However, defining the complexity of a contact is a difficult task as it is a highly abstract concept. While consensus-based data annotation by experienced agents is a possible solution, it is time-consuming and costly. To overcome these challenges, we have developed a novel machine learning approach to define contact complexity. Instead of relying on human annotation, we trained an AI expert model to mimic the behavior of agents and evaluate each contact's complexity based on how the AI expert responds. If the AI expert is uncertain or lacks the skills to comprehend the contact transcript, it is considered a high-complexity contact. Our method has proven to be reliable, scalable, and cost-effective based on the collected data.

cs.LG↗

arXiv2024

Teacher-Student Learning on Complexity in Intelligent Routing

Customer service is often the most time-consuming aspect for e-commerce websites, with each contact typically taking 10-15 minutes. Effectively routing customers to appropriate agents without transfers is therefore crucial for e-commerce success. To this end, we have developed a machine learning framework that predicts the complexity of customer contacts and routes them to appropriate agents accordingly. The framework consists of two parts. First, we train a teacher model to score the complexity of a contact based on the post-contact transcripts. Then, we use the teacher model as a data annotator to provide labels to train a student model that predicts the complexity based on pre-contact data only. Our experiments show that such a framework is successful and can significantly improve customer experience. We also propose a useful metric called complexity AUC that evaluates the effectiveness of customer service at a statistical level.

cs.LG↗

arXiv2024

Uncovering Customer Issues through Topological Natural Language Analysis

E-commerce companies deal with a high volume of customer service requests daily. While a simple annotation system is often used to summarize the topics of customer contacts, thoroughly exploring each specific issue can be challenging. This presents a critical concern, especially during an emerging outbreak where companies must quickly identify and address specific issues. To tackle this challenge, we propose a novel machine learning algorithm that leverages natural language techniques and topological data analysis to monitor emerging and trending customer issues. Our approach involves an end-to-end deep learning framework that simultaneously tags the primary question sentence of each customer's transcript and generates sentence embedding vectors. We then whiten the embedding vectors and use them to construct an undirected graph. From there, we define trending and emerging issues based on the topological properties of each transcript. We have validated our results through various methods and found that they are highly consistent with news sources.

cs.CL↗

arXiv2024

Universal Model in Online Customer Service

Building machine learning models can be a time-consuming process that often takes several months to implement in typical business scenarios. To ensure consistent model performance and account for variations in data distribution, regular retraining is necessary. This paper introduces a solution for improving online customer service in e-commerce by presenting a universal model for predict-ing labels based on customer questions, without requiring training. Our novel approach involves using machine learning techniques to tag customer questions in transcripts and create a repository of questions and corresponding labels. When a customer requests assistance, an information retrieval model searches the repository for similar questions, and statistical analysis is used to predict the corresponding label. By eliminating the need for individual model training and maintenance, our approach reduces both the model development cycle and costs. The repository only requires periodic updating to maintain accuracy.

cs.LG↗

arXiv2018

Atomic-layer-resolved composition and electronic structure of the cuprate Bi$_2$Sr$_2$CaCu$_2$O$_{8+δ}$ from soft x-ray standing-wave photoemission

A major remaining challenge in the superconducting cuprates is the unambiguous differentiation of the composition and electronic structure of the CuO$_2$ layers and those of the intermediate layers. The large c axis for these materials permits employing soft x-ray (930.3 eV) standing wave (SW) excitation in photoemission that yields atomic layer-by-atomic layer depth resolution of these properties. Applying SW photoemission to Bi$_2$Sr$_2$CaCu$_2$O$_{8+δ}$ yields the depth distribution of atomic composition and the layer-resolved densities of states. We detect significant Ca presence in the SrO layers and oxygen bonding to three different cations. The layer-resolved valence electronic structure is found to be strongly influenced by the supermodulation structure--as determined by comparison to DFT calculations, by Ca-Sr intermixing, and by the Cu 3d-3d Coulomb interaction, further clarifying the complex interactions in this prototypical cuprate. Measurements of this type for other quasi-two-dimensional materials with large-c represent a promising future direction.

cond-mat.supr-con↗

arXiv2018

Pressure-induced Frustration of Magnetic Coupling in Elemental Europium

Applying linear response and the magnetic force theorem in correlated density functional theory, the inter-sublattice exchange constants of antiferromagnetic Eu are calculated and found to vanish near the pressure of P$_c$=82 GPa, just where magnetic order is observed experimentally to be lost. The Eu $4f^7$ moment remains unchanged at high pressure, again in agreement with spectroscopic measurements, leaving the picture of perfect frustration of interatomic Ruderman-Kittel-Kasuya-Yoshida couplings in a broad metallic background, leaving a state of electrons strongly exchange coupled to arbitrarily oriented, possibly quasistatic local moments. This strongly frustrated state gives way to superconductivity at T$_c$=1.7K, observed experimentally. These phenomena, and free energy considerations related to correlations, suggest an unusual phase of matter that is discussed within the scenarios of the Doniach Kondo lattice phase diagram, the metallic spin glass class, and itinerant spin liquid or spin gas systems.

cond-mat.str-el↗

arXiv2017

High Surface Conductivity of Fermi Arc Electrons in Weyl semimetals

Weyl semimetals (WSMs), a new type of topological condensed matter, are currently attracting great interest due to their unusual electronic states and intriguing transport properties such as chiral anomaly induced negative magnetoresistance, a semi--quantized anomalous Hall effect and the debated chiral magnetic effect. These systems are close cousins of topological insulators (TIs) which are known for their disorder tolerant surface states. Similarly, WSMs exhibit unique topologically protected Fermi arcs surface states. Here we analyze electron--phonon scattering, a primary source of resistivity in metals at finite temperatures, as a function of the shape of the Fermi arc where we find that the impact on surface transport is significantly dependent on the arc curvature and disappears in the limit of a straight arc. Next, we discuss the effect of strong surface disorder on the resistivity by numerically simulating a tight binding model with the presence of quenched surface vacancies using the Coherent Potential Approximation (CPA) and Kubo--Greenwood formalism. We find that the limit of a straight arc geometry is remarkably disorder tolerant, producing surface conductivity that is a factor of 50 larger of a comparable set up with surface states of TI. Finally, a simulation of the effects of surface vacancies on TaAs is presented, illustrating the disorder tolerance of the topological surface states in a recently discovered WSM material.

cond-mat.mes-hall↗

arXiv2016

New class of 3D topological insulator in double perovskite

We predict a new class of three-dimensional topological insulators (TIs) in which the spin-orbit coupling (SOC) can more effectively generate a large band gap at $Γ$ point. The band gap of conventional TI such as Bi$_2$Se$_3$ is mainly limited by two factors, the strength of SOC and, from electronic structure perspective, the band gap when SOC is absent. While the former is an atomic property, we find that the latter can be minimized in a generic rock-salt lattice model in which a stable crossing of bands {\it at} the Fermi level along with band character inversion occurs for a range of parameters in the absence of SOC. Thus, large-gap TI's or TI's comprised of lighter elements can be expected. In fact, we find by performing first-principle calculations that the model applies to a class of double perovskites A$_2$BiXO$_6$ (A = Ca, Sr, Ba; X = Br, I) and the band gap is predicted up to 0.55 eV. Besides, more detailed calculations considering realistic surface structure indicate that the Dirac cones are robust against the presence of dangling bond at the boundary with a specific termination.

cond-mat.mes-hall↗

arXiv2015

Manipulating Z2 and Chern topological phases in a single material using periodically driving fields

$Z_{2}$ and Chern topological phases such as newly discovered quantum spin Hall and original quantum Hall states hardly both co--exist in a single material due to their contradictory requirement on the time--reversal symmetry (TRS). We show that although the TRS is broken in systems with a periodically driving ac-field, an effective TRS can still be defined provided the ac--field is linearly polarized or certain other conditions are satisfied. The controllable TRS provides us with a route to manipulate $Z_{2} $ and Chern topological phases in a single material by tuning the polarization of the ac--field. To demonstrate the idea, we consider a generic honeycomb lattice model as a benchmark system that is relevant to electronic structures of several monolayered materials. Our calculation shows that not only the transitions between $Z_{2}$ and Chern phases can be induced but also features such as the dispersion of the edge states can be controlled. This opens the possibility of manipulating various topological phases in a single material and can be a promising approach to engineer some new electronic states of matter.

cond-mat.mes-hall↗

arXiv2014

Anisotropic Multipolar Exchange Interactions in Systems with Strong Spin-Orbit Coupling

We introduce a theoretical framework for computaions of anisotropic multipolar exchange interactions found in many spin--orbit coupled magnetic systems and propose a method to extract these coupling constants using a density functional total energy calculation. This method is developed using a multipolar expansion of local density matrices for correlated orbitals that are responsible for magnetic degrees of freedom. Within the mean--field approximation, we show that each coupling constant can be recovered from a series of total energy calculations via what we call the ``pair--flip'' technique. This technique flips the relative phase of a pair of multipoles and computes corresponding total energy cost associated with the given exchange constant. To test it, we apply our method to Uranium Dioxide, which is a system known to have pseudospin $J=1$ superexchange induced dipolar, and superexchange plus spin--lattice induced quadrupolar orderings. Our calculation reveals that the superexchange and spin--lattice contributions to the quadrupolar exchange interactions are about the same order with ferro-- and antiferro--magnetic contributions, respectively. This highlights a competition rather than a cooperation between them. Our method could be a promising tool to explore magnetic properties of rare--earth compounds and hidden--order materials.

cond-mat.str-el↗

arXiv2013

Calculation of Multipolar Exchange Interactions in Spin-Orbital Coupled Systems

A new method of computing multipolar exchange interaction in spin-orbit coupled systems is developed using multipolar tensor expansion of the density matrix in LDA+U electronic structure calculation. Within mean-field approximation, exchange constants can be mapped into a series of total energy calculations by pair-flip technique. Application to Uranium dioxide shows an antiferromagnetic superexchange coupling in dipoles but ferromagnetic in quadrupoles which is very different from past studies. Further calculation of spin-lattice interaction indicates it is of the same order with superexchange and characterizes the overall behavior of quadrupolar part as a competition between them.

cond-mat.str-el↗