Search

Recommender systems are pivotal in enhancing user experiences across various web applications by analyzing the complicated relationships between users and items. Knowledge graphs(KGs) have been widely used to enhance the performance of recommender systems. However, KGs are known to be noisy and incomplete, which are hard to provide reliable explanations for recommendation results. An explainable recommender system is crucial for the product development and subsequent decision-making. To address these challenges, we introduce a novel recommender that synergies Large Language Models (LLMs) and KGs to enhance the recommendation and provide interpretable results. Specifically, we first harness the power of LLMs to augment KG reconstruction. LLMs comprehend and decompose user reviews into new triples that are added into KG. In this way, we can enrich KGs with explainable paths that express user preferences. To enhance the recommendation on augmented KGs, we introduce a novel subgraph reasoning module that effectively measures the importance of nodes and discovers reasoning for recommendation. Finally, these reasoning paths are fed into the LLMs to generate interpretable explanations of the recommendation results. Our approach significantly enhances both the effectiveness and interpretability of recommender systems, especially in cross-selling scenarios where traditional methods falter. The effectiveness of our approach has been rigorously tested on four open real-world datasets, with our methods demonstrating a superior performance over contemporary state-of-the-art techniques by an average improvement of 12%. The application of our model in a multinational engineering and technology company cross-selling recommendation system further underscores its practical utility and potential to redefine recommendation practices through improved accuracy and user trust.

In fifth generation (5G) wireless cellular networks, millimeter wave spectrum opens room for several potential improvements in throughput, reliability, latency, among other aspects. However, it also brings challenges, such as a higher influence of blockage which may significantly limit the coverage. In this context, network-controlled repeaters (NCRs) are network nodes with low complexity that represent a technique to overcome coverage problems. In this paper, we introduce the NCR concept and study its performance gains and deployment options. Particularly, presenting the main specifications of NCR as agreed in 3rd generation partnership project (3GPP) Rel-18, we analyze different NCR deployments in an urban scenario and compare its performance with alternative deployments. As demonstrated, with a proper network planning and beamforming design, NCR is an attractive solution to cover blind spots the base stations (BSs) may have., Comment: Submmited to IEEE Communications Standards Magazine

Learning adversarial examples can be formulated as an optimization problem of maximizing the loss function with some box-constraints. However, for solving this induced optimization problem, the state-of-the-art gradient-based methods such as FGSM, I-FGSM and MI-FGSM look different from their original methods especially in updating the direction, which makes it difficult to understand them and then leaves some theoretical issues to be addressed in viewpoint of optimization. In this paper, from the perspective of adapting step-size, we provide a unified theoretical interpretation of these gradient-based adversarial learning methods. We show that each of these algorithms is in fact a specific reformulation of their original gradient methods but using the step-size rules with only current gradient information. Motivated by such analysis, we present a broad class of adaptive gradient-based algorithms based on the regular gradient methods, in which the step-size strategy utilizing information of the accumulated gradients is integrated. Such adaptive step-size strategies directly normalize the scale of the gradients rather than use some empirical operations. The important benefit is that convergence for the iterative algorithms is guaranteed and then the whole optimization process can be stabilized. The experiments demonstrate that our AdaI-FGM consistently outperforms I-FGSM and AdaMI-FGM remains competitive with MI-FGSM for black-box attacks.

Network-controlled repeater (NCR) has been recently considered as a study-item in 3GPP Release 18, and the discussions are continuing in a work-item. In this paper, we introduce the concept of NCRs, as a possible low-complexity device to support for network densification and compare the performance of the NCRs with those achieved by reconfigurable intelligent surfaces (RISs). The results are presented for the cases with different beamforming methods and hardware impairment models of the RIS. Moreover, we introduce the objectives of the 3GPP Release 18 NCR work-item and study the effect of different parameters on the performance of NCR-assisted networks. As we show, with a proper deployment, the presence of NCRs/RISs can improve the network performance considerably., Comment: 7 pages, 7 figures, submitted to potential IEEE publication

Newton's third law is one of the most important concepts learned early in introductory mechanics courses; however, ample studies have documented a wide range of students' misconceptions and fragmented understandings of this concept that are difficult to change through traditional instruction. This research develops a conceptual framework model to investigate students' understanding of Newton's third law through the knowledge integration perspective. The conceptual framework is established with a central idea emphasizing forces as the outcomes of an interaction instead of using the common action-reaction language. Additional studies to apply this conceptual framework in assessment and instruction will follow.

The dissolution of liquid nanodroplets is a crucial step in many applied processes, such as separation and dispersion in food industry, crystal formation of pharmaceutical products, concentrating and analysis in medical diagnosis, and drug delivery in aerosols. In this work, using both experiments and numerical simulations, we \textit{quantitatively} study the dissolution dynamics of femtoliter surface droplets in a highly ordered array under a uniform flow. Our results show that the dissolution of femoliter droplets strongly depends on their spatial positions relative to the flow direction, drop-to-drop spacing in the array, and the imposed flow rate. In some particular case, the droplet at the edge of the array can dissolve about 30% faster than the ones located near the centre. The dissolution rate of the droplet increases by 60% as the inter-droplet spacing is increased from 2.5 $\mu$m to 20 $\mu$m. Moreover, the droplets close to the front of flow commence to shrink earlier than those droplets in the center of the array. The average dissolution rate is faster for faster flow. As a result, the dissolution time $T_{i}$ decreases with the Reynolds number Re of the flow as $T_{i}\propto Re^{-3/4}$. The experimental results are in good agreement with numerical simulations where the advection-diffusion equation for the concentration field is solved and the concentration gradient on the surface of the drop is computed. The findings suggest potential approaches to manipulate nanodroplet sizes in droplet arrays simply by dissolution controlled by an external flow. The obtained droplets with varying curvatures may serve as templates for generating multifocal microlens in one array.

Nanodroplets on a solid surface (i.e. surface nanodroplets) have practical implications for high-throughput chemical and biological analysis, lubrications, lab-on-chip devices, and near-field imaging techniques. Oil nanodroplets can be produced on a solid-liquid interface in a simple step of solvent exchange in which a good solvent of oil is displaced by a poor solvent. In this work, we experimentally and theoretically investigate the formation of nanodroplets by the solvent exchange process under well-controlled flow conditions. We find that the contact angle of the nanodroplets is independent of the flow condition. However, there are significant effects from the flow rate and the flow geometry on the droplet size. We develop a theoretical framework to account for these effects. The main idea is that the droplet nuclei are exposed to an oil oversaturation pulse during the exchange process. The analysis gives that the volume of the nanodroplets increases with the Peclet number $Pe$ of the flow as $\propto Pe^{3/4}$, which is in good agreement with our experimental results. In addition, at fixed flow rate and thus fixed Peclet number, larger and less homogeneously distributed droplets formed at less narrow channels, due to convection effects originating from the density difference between the two solutions of the solvent exchange. The understanding from this work provides valuable guidelines for producing surface nanodroplets with desired sizes by controlling the flow conditions., Comment: 24 pages, 5 figures

We propose and demonstrate a computational imaging technique that uses structured illumination based on a two-dimensional discrete cosine transform to perform imaging with a single-pixel detector. A scene is illuminated by a projector with two sets of orthogonal patterns, then by applying an inverse cosine transform to the spectra obtained from the single-pixel detector a full-color image is retrieved. This technique can retrieve an image from sub-Nyquist measurements, and the background noise is easily canceled to give excellent image quality. Moreover, the experimental setup is very simple., Comment: 7 pages,4 figures

The structural flexibility of nucleic acids plays a key role in many fundamental life processes, such as gene replication and expression, DNA-protein recognition, and gene regulation. To obtain a thorough understanding of nucleic acid flexibility, extensive studies have been performed using various experimental methods and theoretical models. In this review, we will introduce the progress that has been made in understanding the flexibility of nucleic acids including DNAs and RNAs, and will emphasize the experimental findings and the effects of salt, temperature, and sequence. Finally, we will discuss the major unanswered questions in understanding the flexibility of nucleic acids., Comment: 24 pages, 3 tables

Flexibility of short DNA helices is important for the biological functions such as nucleosome formation and DNA-protein recognition. Recent experiments suggest that short DNAs of tens of base pairs (bps) may have apparently higher flexibility than those of kilo bps, while there is still the debate on such high flexibility. In the present work, we have studied the flexibility of short DNAs with finite-length of 5 to 50 bps by the all-atomistic molecular dynamics simulations and Monte Carlo simulations with the worm-like chain model. Our microscopic analyses reveal that short DNAs have apparently high flexibility which is attributed to the significantly strong bending and stretching flexibilities of ~6 bps at each helix end. Correspondingly, the apparent persistence length lp of short DNAs increases gradually from ~29nm to ~45nm as DNA length increases from 10 to 50 bps, in accordance with the available experimental data. Our further analyses show that the short DNAs with excluding ~6 bps at each helix end have the similar flexibility with those of kilo bps and can be described by the worm-like chain model with lp~50nm., Comment: 32 pages, 9 figures,Supplementary Material

Cognitive conflict is well recognized as an important factor in conceptual change and is widely used in developing inquiry-based curricula. However, cognitive conflict can also contribute to student anxiety during learning, which can have both positive and negative impacts on students' motivation and learning achievement. Therefore, instructors need to be informed of the impacts of introducing cognitive conflicts during teaching. To get this information, teachers need a practical instrument that can help them identify the existence and features of cognitive conflict introduced by the instruction and the resulting anxiety. Based on the literature on studies of cognitive conflict and student anxiety, a quantitative instrument, the In-class Conflict and Anxiety Recognition Evaluation (iCARE), was developed and used to monitor the status of students' cognitive conflict and anxiety in the Physics by Inquiry (PBI) classes. This paper introduces this instrument and discusses the types of information that can be measured. Research and pedagogical values of this instrument are also discussed., Comment: 52 pages

Item Response Theory (IRT) is a popular assessment method used in education measurement, which builds on an assumption of a probability framework connecting students' innate ability and their actual performances on test items. The model transforms students' raw test scores through a nonlinear regression process into a scaled proficiency rating, which can be used to compare results obtained with different test questions. IRT also provides a theoretical approach to address ceiling effect and guessing. We applied IRT to analyze the Force Concept Inventory (FCI). The data was collected from 2802 students taking intro level mechanics courses at The Ohio State University. The data was analyzed with a 3-parameter item response model for multiple choice questions. We describe the procedures of the analysis and discuss the results and the interpretations. The analysis outcomes are compiled to provide a detailed IRT measurement metric of the FCI, which can be easily referenced and used by teachers and researchers for a range of assessment applications., Comment: 18 pages, 5 figures

Student content knowledge and general reasoning abilities are two important areas in education practice and research. However, there hasn't been much work in physics education that clearly documents the possible interactions between content learning and the development of general reasoning abilities. In this paper, we report one study of a systematic research to investigate the possible interactions between students' learning of physics content knowledge and the development of general scientific reasoning abilities. Specifically, this study seeks to answer the research question of whether and to what extent content learning may affect the development of general reasoning abilities. College entrance testing data of freshman college students in both USA and China were collected using three standardized tests, FCI, BEMA, and Lawson's Classroom Test of Scientific Reasoning (Lawson Test). The results suggest that years of rigorous training of physics knowledge in middle and high schools have made significant impact on Chinese students' ability in solving physics problems, while such training doesn't seem to have direct effects on their general ability in scientific reasoning, which was measured to be at the same level as that of the students in USA. Details of the curriculum structures in the education systems of USA and China are also compared to provide a basis for interpreting the assessment data., Comment: To appear in AJP

Multiple-choice tests such as the Force Concept Inventory (FCI) provide useful instruments to probe the distribution of student difficulties on a large scale. However, traditional analysis often relies solely on scores (number of students giving the correct answer). This ignores what can be significant and important information: the distribution of wrong answers given by the class. In this paper a new method is introduced called "concentration analysis" to measure how students' responses on multiple-choice questions are distributed. This information can be used to study if the students have common incorrect models or if the question is effective in detecting student models. When combined with information obtained from qualitative research, the method can be used to identify cleanly what FCI results are saying about student knowledge. (Contains 19 references, 9 figures, and 8 tables.) (Author/YDS)

Pre-post testing is a commonly used method in physics education community for evaluating students' achievement and or the effectiveness of teaching through a specific period of instruction. A popular method to analyze pre-post testing results is the normalized gain first brought to the physics education community in wide use by R.R. Hake. This paper presents a measurement based probabilistic model of the dynamic process of learning that explains the experimentally observed features of the normalized gain. In Hake's study with thousands of students' pre-post testing results, he observed that on average 48 courses employing "interactive engagement" types of instruction achieved average normalized gains about two standard deviations greater than did 14 courses subjected to traditional instruction. For all courses the average normalized gains had a very low correlation +0.02 with average pretest scores. This feature of the normalized gain has allowed researchers to investigate the effectiveness of instruction using data collected from classes with widely different average pretest scores. However, the question of why the average normalized gain has this feature and to what extent this feature is generally present is not well understood. In addition, there have been debates as to what the normalized gain actually measures, and concerns that it lacks a probability framework that undergirds psychometric methods such as Item Response Theory (IRT). The present model leads to an explanation of the observed features of the normalized gain, connects to other models such as IRT, and shows that the normalized gain does have a probability framework but one different from that emphasized by IRT., Comment: Theoretical Models of Education Measurement

We report on a multi-year, multi-institution study to investigate student reasoning about energy in the context of quantum tunnelling. We use ungraded surveys, graded examination questions, individual clinical interviews, and multiple-choice exams to build a picture of the types of responses that students typically give. We find that two descriptions of tunnelling through a square barrier are particularly common. Students often state that tunnelling particles lose energy while tunnelling. When sketching wave functions, students also show a shift in the axis of oscillation, as if the height of the axis of oscillation indicated the energy of the particle. We find inconsistencies between students' conceptual, mathematical, and graphical models of quantum tunnelling. As part of a curriculum in quantum physics, we have developed instructional materials to help students develop a more robust and less inconsistent picture of tunnelling, and present data suggesting that we have succeeded in doing so., Comment: Originally submitted to the European Journal of Physics on 2005 Feb 10. Pages: 14. References: 11. Figures: 9. Tables: 1. Resubmitted May 18 with revisions that include an appendix with the curriculum materials discussed in the paper (4 page small group UW-style tutorial)

We are trying to identify resources students are using to reason in quantum mechanics. In this process we realize students must have not only the right resources available but sophisticated for evaluating and controlling their thought processes. We will discuss examples from student interviews to illustrate our point., Comment: 4 pages, No figures, Submitted to the Physics Education Research Conference Proceedings, 2002

In this paper we present a method of modeling and analysis that permits the extraction and quantitative display of detailed information about the effects of instruction on a class's knowledge. The method relies on a congitive model that represents student thinking in terms of mental models. Students frequently fail to recognize relevant conditions that lead to appropriate uses of their models. As a result they can use multiple models inconsistently. Once the most common mental models have been determined by qualitative research, they can be mapping onto a multiple choice test. Model analysis permits the interpretation of such a situation. We illustrate the use of our method by analyzing results from the FCI., Comment: 25 pages, submitted to Cognition and Instruction, March 2001