Random Reordering Research Articles

Sparse Enhanced Network: An Adversarial Generation Method for Robust Augmentation in Sequential Recommendation

Sequential Recommendation plays a significant role in daily recommendation systems, such as e-commerce platforms like Amazon and Taobao. However, even with the advent of large models, these platforms often face sparse issues in the historical browsing records of individual users due to new users joining or the introduction of new products. As a result, existing sequence recommendation algorithms may not perform well. To address this, sequence-based data augmentation methods have garnered attention. Existing sequence enhancement methods typically rely on augmenting existing data, employing techniques like cropping, masking prediction, random reordering, and random replacement of the original sequence. While these methods have shown improvements, they often overlook the exploration of the deep embedding space of the sequence. To tackle these challenges, we propose a Sparse Enhanced Network (SparseEnNet), which is a robust adversarial generation method. SparseEnNet aims to fully explore the hidden space in sequence recommendation, generating more robust enhanced items. Additionally, we adopt an adversarial generation method, allowing the model to differentiate between data augmentation categories and achieve better prediction performance for the next item in the sequence. Experiments have demonstrated that our method achieves a remarkable 4-14% improvement over existing methods when evaluated on the real-world datasets. (https://github.com/junyachen/SparseEnNet)

Correction: Random reordering in SOR-type methods

Correction: Random reordering in SOR-type methods

Solving the Olum 2 cipher: a new approach to cryptanalysis of transposition ciphers

Olum 2 is one of two ciphers created more than 75 years ago by mathematician Paul Olum to challenge his Manhattan Project officemate, physicist Richard Feynman. In this manuscript, I describe the first successful decryption of Olum 2 using a novel approach to cryptanalysis of transposition ciphers. To decrypt Olum 2, I generated the bigrams and trigrams for all possible transposition intervals. I then identified transposition intervals with multiple bigrams and trigrams that occur frequently in English. I calculated the ratios of their English frequencies to the frequencies of bigrams and trigrams generated by a random reordering of the ciphertext. This enabled me to identify letter sequences with the highest probability of being true cipher message components rather than occurring by chance. In Olum 2, Professor Olum divided the message into sections of thirty-five letters and applied a rotating key to change the order of transposition for each successive section. His strategy not only confounded Professor Feynman but also proved impervious to several decryption programs in use today that assume a uniform transposition has been applied throughout the cipher. The decryption methods described in this manuscript can assist in the decryption of other ciphers employing a variety of transposition methods.

Stable and efficient retrospective 4D-MRI using non-uniformly distributed quasi-random numbers

The purpose of this work is the development of a robust and reliable three-dimensional (3D) Cartesian imaging technique for fast and flexible retrospective 4D abdominal MRI during free breathing. To this end, a non-uniform quasi random (NU-QR) reordering of the phase encoding (ky–kz) lines was incorporated into 3D Cartesian acquisition. The proposed sampling scheme allocates more phase encoding points near the k-space origin while reducing the sampling density in the outer part of the k-space. Respiratory self-gating in combination with SPIRiT-reconstruction is used for the reconstruction of abdominal data sets in different respiratory phases (4D-MRI). Six volunteers and three patients were examined at 1.5 T during free breathing. Additionally, data sets with conventional two-dimensional (2D) linear and 2D quasi random phase encoding order were acquired for the volunteers for comparison. A quantitative evaluation of image quality versus scan times (from 70 s to 626 s) for the given sampling schemes was obtained by calculating the normalized mutual information (NMI) for all volunteers. Motion estimation was accomplished by calculating the maximum derivative of a signal intensity profile of a transition (e.g. tumor or diaphragm).The 2D non-uniform quasi-random distribution of phase encoding lines in Cartesian 3D MRI yields more efficient undersampling patterns for parallel imaging compared to conventional uniform quasi-random and linear sampling. Median NMI values of NU-QR sampling are the highest for all scan times. Therefore, within the same scan time 4D imaging could be performed with improved image quality.The proposed method allows for the reconstruction of motion artifact reduced 4D data sets with isotropic spatial resolution of 2.1 × 2.1 × 2.1 mm3 in a short scan time, e.g. 10 respiratory phases in only 3 min. Cranio-caudal tumor displacements between 23 and 46 mm could be observed.NU-QR sampling enables for stable 4D-MRI with high temporal and spatial resolution within short scan time for visualization of organ or tumor motion during free breathing. Further studies, e.g. the application of the method for radiotherapy planning are needed to investigate the clinical applicability and diagnostic value of the approach.

Mitigation of Sampling Errors in VCO-Based ADCs

Voltage-controlled-oscillator-based analog-to-digital converter (ADC) is a scaling-friendly architecture to build ADCs in fine-feature complimentary metal-oxide-semiconductor processes. Lending itself to an implementation with digital components, such a converter enables design automation with existing digital CAD hence reducing design and porting costs compared with a custom design flow. However, robust architectures and circuit techniques that reduce the dependence of performance on component accuracy are required to achieve good performance while designing converters with low accuracy components like standard cells in deeply-scaled processes. This paper investigates errors resulting from the sampling of a fast switching multi-phase ring oscillator output. A scheme employing ones-counters is proposed to encode the sampled ring oscillator code into a binary representation, which is resilient to a class of sampling induced errors modeled by the temporal reordering of the transitions in the ring. In addition to correcting errors caused by deterministic reordering, proposed encoding suppresses conversion errors in the presence of arbitrary reordering patterns that may result from automatic place-and-route in wire-delay dominated processes. The error suppression capability of the encoding is demonstrated using MATLAB simulation. The proposed encoder reduces the error caused by the random reordering of six subsequent bits in the sampled signal from 31 to 2 LSBs for a 31-stage oscillator.

Random reordering in SOR-type methods

When iteratively solving linear systems $$By=b$$By=b with Hermitian positive semi-definite B, and in particular when solving least-squares problems for $$Ax=b$$Ax=b by reformulating them as $$AA^*y=b$$AAźy=b, it is often observed that SOR type methods (Gauβ-Seidel, Kaczmarz) perform suboptimally for the given equation ordering, and that random reordering improves the situation on average. This paper is an attempt to provide some additional theoretical support for this phenomenon. We show error bounds for two randomized versions, called shuffled and preshuffled SOR, that improve asymptotically upon the best known bounds for SOR with cyclic ordering. Our results are based on studying the behavior of the triangular truncation of Hermitian matrices with respect to their permutations.

Note on the hypergeometric distribution as an invalidation test for binary forecasts

One mechanism giving rise to the hypergeometric distribution is the number of matches in a random reordering of a sequence of forecasts of binary events. This provides a simple means of invalidating a time series of binary forecasts if in fact the forecasting method has no skill. The hypergeometric distribution has a long history of application in this context but appears not to have a high profile in the environmental sciences. Attention is drawn to the utility of the distribution as a simple nonparametric test of the null hypothesis of no skill when forecasting binary environmental outcomes.

Seasonality in live fish movements and its effects on epidemic dynamics

Live fish movements between salmon farms risk spreading pathogens, and movements between freshwater farms (FW–FW) or from freshwater to seawater (FW–SW) show clear seasonality. In this study, we quantify the effects of seasonality of live fish movements on epidemic dynamics, using a network model populated with data from live fish movements between Scottish salmon farms from 2002 to 2004. We used three types of networks: A) timing and pair-wise movements between farms were as observed; B) as network A, but with a random reordering of FW–FW and FW–SW movements; and C) number of movements were kept the same as in the data, but connection between nodes was random. We compared the time-course of simulated epidemics in a stochastic model for all three networks. We showed that seasonality had the strongest effect in networks B and C, especially when local transmission was high, and this effect was stronger in SW farms compared with FW farms.

Assessing the Efficacy of the Fundamentals of Research and Career Development Course Overseas

As the Fundamentals of Research and Career Development Course (FRCDC) is conducted internationally, questions have arisen regarding the cultural appropriateness of the United States (US) course. We therefore assessed the US-based teaching methodology during the FRCDC in Abuja, Nigeria. We hypothesized that the US-based instructional methods would be effective. Twenty questions were distributed to attendees of the FRCDC prior to commencement. The same 20 questions were administered at the conclusion of the course after random reordering. Differences between the pre- and post-test results were assessed for normalcy and compared using the paired t-test. There were 89 attendees, of whom 60 completed the pre-test and 77 completed the post-test. The pre-test group answered 12.3 ± 2.6 questions correctly, which improved to 15.0 ± 2.6 in the post-test group (P < 0.001). On the pre-test, the least common correct answers were for questions regarding type 1 and 2 error (16.7% correct), the definition of health services and outcomes research (26.7%), and how to best address missing data (26.7%). On the post-test, the questions with the least common correct answers were regarding the definition of health services and outcomes research (35%), and the components of an NIH grant (37.7%). Our results suggest that the FRCDC in Nigeria as given by US faculty has short-term efficacy. Attendees were able to improve their scores despite the cultural differences between them and the lecturers. Our next goal will be to demonstrate long-term efficacy at future courses in the region using similar questionnaire strategies.

An invalidation test for predictive models

The standard means of establishing predictive ability in hydrological models is by finding how well predictions match independent validation data. This matching may not be particularly good in some situations such as seasonal flow forecasting and the question arises as to whether a given model has any predictive capacity. A model-independent significance test of the presence of predictive ability is proposed through random permutations of the predicted values. The null hypothesis of no model predictive ability is accepted if there is a sufficiently high probability that a random reordering of the predicted values will yield a better fit to the validation data. The test can achieve significance even with poor model predictions and its value is for invalidating bad models rather than verifying good models as suitable for application. Some preliminary applications suggest that test outcomes will often be similar at the 0.05 level for standard fit measures using absolute or squared residuals. In addition to hydrological application, the test may also find use as a base quality control measure for predictive models generally.

The multifractal analysis of the occupation measure of a Lévy process

We introduce the results on the multifractal structure of the occupation measures of a Brownian Motion, a stable process, a general subordinator and a stochastic process derived from random reordering of the Cantor set. We also introduced an interesting and powerful technique to investigate the multifractal spectrum.

Packet reordering is not pathological network behavior

It is a widely held belief that packet reordering in the Internet is a pathological behavior, or more precisely, that it is an uncommon behavior caused by incorrect or malfunctioning network components. Some studies of Internet traffic have reported seeing occasional packet reordering events and ascribed these events to route fluttering, router pauses or simply to broken equipment. We have found, however, that parallelism in Internet components and links is causing packet reordering under normal operation and that the incidence of packet reordering appears to be substantially higher than previously reported. More importantly, we observe that in the presence of massive packet reordering transmission control protocol (TCP) performance can be profoundly effected. Perhaps the most disturbing observation about TCP's behavior is that large scale and largely random reordering on the part of the network can lead to self-reinforcingly poor performance from TCP.

A BASIC program for the generation of Latin squares

I~ experimental research, it is oftendesirableto expose subjects to more than one stimulus or experimentalconditionto provide for the greaterpowerinherent in repeated measures designs, the economical use of available subjects, or the within-subjects test thatconceptual hypotheses may specifically require. A potential problemin thesedesignsis thepossibility of carryovereffects or the influence of a particular order of presentation. One popular counterbalancing strategy involves the selection of k orders from all possible orders by construction of a Latinsquare, with subjects assigned randomly to one of the k orders. The advantage of the Latinsquareis thata relatively small number of orders is required (i.e., k) and that the orders are selected in such a way as to guarantee the existence of ce~in features. Specifically, each stimulus appears once m every ordinal position, and the set selected is cho~en randomly from among all possible Latin squares of size k. Furthermore, specific statistical analyses have been developed for Latinsquaredesigns, whichallowfor the testingof certain order effects (see, e.g., Kirk, 1968; Myers, 1972; Winer, 1971). Cons~ction of a Latinsquarerequires four basic steps: (1) selection of a standard square of size k, (2) random reordering of the columns of the standard square, ~3) random reordering of the rows of the square result109 from steps I and 2, and (4) random shufflingof the numerical indicators (e.g., Myers, 1972). The resulting rows (or columns) of the final square constitute the k orders, to which each subject in the study is randomly assigned. . The ~jor drawback.to usingthe Latinsquareapproach IS that It can be very tIme-consuming to implement. As k increases, the number of standard squares increases dramatically, and one square must be chosen at random as the startingpoint for the reorderingof rows, columns, and indicators (eachstandard squaregivesrise to a unique subset of all possible Latin squares). For example, although there isonly I standard squarefor k=2 andk=3, and only 4 standard squares for k=4, there are 56 standard squares for k=5, 9,408 standard squares for k=6,

Mixed excitation for speech II: Naturalness of period‐by‐period reordered voiced fricatives

Previous attempts at synthesizing voiced fricatives have failed to yield acceptably natural‐sounding segments. One possible reason for this failure is that the synthesis models have been global, whereas some important characteristics of these segments may reside in their period‐by‐period structure. Our goal in this study was therefore to compare the auditory quality of digitally manipulated /v, ð, z, and ȝ/ segments, in various V‐V contexts. High‐quality tokens were digitized at 24 kHz, and individual pitch periods were marked by hand using interactive software. Comparison stimuli were then created via the following manipulations: (1) reordering of alternate periods; (2) reordering triples of periods; (3) replacement of all odd‐ (even‐) numbered periods by their even‐ (odd‐) numbered neighbors; (4) similar replacements modulo 3; (5) random reordering of periods. Furthermore, these manipulations were sometimes restricted to the onset, steady state, or offset portions of the segments. These digitally spliced segments were then presented to listeners for discrimination, naturalness, and likeness judgments. Results of these comparisons and their implications for synthesis of natural‐sounding voiced fricatives will be discussed.

Attitudes toward Abortion in Thailand: A Survey of Senior Medical Students

A majority of senior Thai medical students favored induced abortion in many circumstances not now considered legal. While 20 percent would not perform abortions for any reason, many of these would refer cases to colleagues. Independent variables consistently and significantly influencing attitudes conservatively were: female sex, certain medical schools attended, larger ideal family size, intention not to teach in a medical school as a career, having a father in unskilled or semiskilled or work, never having witnessed an illegal abortion, and never having been asked to find an abortion practitioner for a friend. A random reordering of the questions did not affect reports of attitudes toward or willingness to perform induced abortion.

Order Constraints in American Sign Language: The Effects of Structure on Judgments of Meaningfulness and on Immediate Recall of Anomalous Sign Sequences

Forty anomalous sign sequences were ranked as to their meaningfulness by three judges working independently, and the same forty sequences were presented to seventeen deaf subjects via videotape in a task measuring immediate recall. Half of the sequences preserved the structure of original model ASL sequences and half of the sequences were the result of a random reordering of the original sign sequence. Twenty of the sentences were five signs long and twenty were six signs long. Random resequencing of the signs affected judgments of meaningfulness in both five- and six-sign sequences, and resequencing of the six-sign sequences affected both meaningfulness and immediate recall. It was concluded that order constraints are present in American Sign Language and that the sequence of signs may affect both semantic features and psychological processing of signed messages.