This article presents hardware-efficient harmonics cancellation techniques using digital predistortion (DPD) for low-distortion sinusoidal signal generation, targeted for the dynamic characteristic testing of high-resolution 14-bit and 16-bit 1 MS/s analog-to-digital converters (ADCs) in VLSI testing environments. For 14-bit ADC testing, a single 16-bit digital-to-analog converter (DAC) with a simplified DPD circuit is demonstrated to achieve a second-order harmonic distortion (HD2) level of −92dBc. For 16-bit ADC testing, where higher resolution is required, we propose a novel dual-DAC architecture utilizing a dual-range synthesis method to enhance effective resolution and suppress harmonics up to the fifth-order (HD2–HD5). Furthermore, to minimize the hardware overhead in practical implementations, we introduce optimized field-programmable gate array (FPGA) design strategies, including time-division multiplexing (TDM) and compressed sine look-up tables (LUTs), achieving a 96.8% reduction in memory usage. The parameters are extracted via a single-step, deterministic calibration flow, ensuring high testing throughput without iterative optimization. Experimental results show HD2–HD5 levels down to −120 dBc, providing a hardware-efficient solution for high-precision built-out self-test (BOST) and future system-on-chip (SoC) integration.

With the rapid growth of scientific publications, researchers need to spend more time and effort searching for papers that align with their research interests. To address this challenge, paper recommendation systems have been developed to help researchers in effectively identifying relevant paper. One of the leading approaches to paper recommendation is content-based filtering method. Traditional content-based filtering methods recommend relevant papers to users based on the overall similarity of papers. However, these approaches do not take into account the information seeking behaviors that users commonly employ when searching for literature. Such behaviors include not only evaluating the overall similarity among papers, but also focusing on specific sections, such as the method section, to ensure that the approach aligns with the user’s interests. In this paper, we propose a content-based filtering recommendation method that takes this information seeking behavior into account. Specifically, in addition to considering the overall content of a paper, our approach also takes into account three specific sections (background, method, and results) and assigns weights to them to better reflect user preferences. We conduct offline evaluations on the publicly available DBLP dataset, and the results demonstrate that the proposed method outperforms six baseline methods in terms of precision, recall, F1-score, MRR, and MAP.

Abstract This paper presents a new approach to dense retrieval across multiple modalities, emphasizing the integration of images and sensor data. Traditional cross-modal retrieval techniques face significant challenges, particularly in processing non-linguistic modalities and creating effective training datasets. To address these issues, we propose a method that uses a shared vector space, optimized with contrastive loss, to enable efficient and accurate retrieval across diverse modalities. A key innovation of our approach is the introduction of a temporal closeness metric, which evaluates the relationship between data points based on their timestamps. This metric helps automatically extract positive and hard negative samples related to the query, improving the training process and enhancing the retrieval model’s performance. We validate our approach using the Lifelog Search Challenge 2024 (LSC’24) dataset, one of the largest multi-modal datasets, including non-linguistic data such as egocentric images, heart rate, and location information. Our evaluation shows that incorporating temporal closeness into the dense retrieval process significantly improves retrieval accuracy and robustness in real-world, multi-modal scenarios. This paper’s contributions include developing a novel dense retrieval framework, introducing the temporal closeness metric, and successfully applying these innovations to a comprehensive multi-modal dataset.

Humans make decisions based on their internal value function, and its shape is known to be distorted and biased around a point, which the research community of behavior economics refers to as the reference point. People intensify activities that come to lie within the reach of their reference point, and abstain from acts that would incur losses once they've crossed the point. However, the impact of past experiences on decision making around the reference point has not been well studied. By analyzing a long series of user-level decisions gathered from a competitive programming website, we find that history has a clear impact on user's decision making around the reference point. Past experiences can strengthen, and sometimes weaken, the decision bias around the reference point. Experiences of past difficulties can strengthen the tendency towards loss aversion after achieving the reference point. When a person crosses a reference point for the first time, the cognitive decision bias is significant. However, repeating this crossing gradually weakens the effect. We also show the value of our insights in the task of predicting user behavior. Prediction models incorporating our insights may be used for motivating people to remain more active.

This paper presents a physically unclonable function (PUF) using flash time-to-digital converter (TDC) with linearity self-calibration. The proposed PUF utilizes that the variation of delay of delay elements of TDC is unique to the device and unclonable. The proposed PUF is constructed using the flash TDC with linearity self-calibration using histogram method. With the linearity self-calibration operation, variation of delay elements is estimated. The response output of the PUF is calculated using the estimated variation and the challenge inputs. The proposed PUF is a simple digital circuit consisting of basic digital elements. It is easy to design and implement to both SoC and FPGA. It can be used as not only strong PUF but also as a TDC with fine linearity. The experimental results with Artix7 FPGA show that the intra-chip variation is 8.9 % and the inter-chip variation is 46.9 %. The probability of the correct identification is 99.8 %. Extra resources to construct the proposed PUF are 33.7 % of the resources of the TDC with linearity self-calibration.

This paper describes two harmonics cancellation techniques using digital predistortion (DPD) for low-distortion sinusoidal signal generation with direct digital synthesizer; it is targeted for the dynamic characteristic testing of 14-bit and 16-bit 1MS/s ADCs and their testing sinusoidal wave frequencies are around 100kHz. (i) The first one is an analog-intensive DPD method for the 16-bit ADC testing. The HD2 and HD3 of the original sinusoidal signal are measured with FFT method, and then their 180-degree phase-shifted signals are generated with an auxiliary DAC and added to the original signal to cancel the HD2 and HD3. Our experiments show that −120dBc of HD2 and HD3 can be achieved. (ii) The second one is a digital-intensive DPD method for the 14-bit ADC testing. The digital data of the 180-degree phase-shifted signals of the measured HD2 and HD3 are added to the original digital data of the direct digital synthesizer. The circuit is simple, and our experiments show that −92dBc of HD2 can be achieved. Their details of analysis, simulation and experimental results are shown.

Data often accumulates in tabular format with many attribute items, and prediction using machine learning adds value to data for business. However, studies on machine learning for tabular data only input attribute values, which reduces accuracy. Therefore, we propose an inference method that inputs attribute values and values from aggregated tabular data that has varying attribute values for each attribute item. In an experiment, we compared our proposed method with AutoGluon-Tabular using AutoML benchmark datasets. Our proposed method achieved the highest accuracy for 21 out of 39 datasets.

In LSI testing, equivalent time sampling techniques are frequently used because the input signals to the device under test and the sampling clock are controllable; when the repetitive input signals are applied to the analog device under test, its output signals can be also repetitive. In this paper, we investigate an efficient waveform acquisition method with the equivalent-sampling using the metallic ratio of the sampling frequency and the input frequency, which is expected to be used for on-line, short-time and simple analog/RF/mixed-signal IC testing.

This paper describes that the mature ADC testing method with a simple test system using the incoherent sampling and the standard algorithm of windowing and FFT with 4Kpoint data can measure the SINAD of our target 12-bit SAR ADC accurately by proper setting of the input and sampling frequencies, which is industry-friendly. We show the input sinusoidal signal and sampling clock frequency relationship for accurate testing of the ADC dynamic characteristics with an incoherent sampling method using a flat-top window. We have clarified the measured SINAD accuracy of the input signal frequency dependency for a fixed sampling frequency, a specified resolution of the ADC under test and a given number of FFT points (data samples) in the incoherent sampling environment. Mature technology combinations with their optimal usage and without advanced methods can lead to the low-cost high-quality testing of the ADC, which can be well accepted in industry. Their analysis, simulation and experimental results are shown.

Many people share their daily events and opinions on Twitter. Some tweets are beneficial and others are related to such aspects of a user’s real-life as eating, traffic conditions, and weather. In this paper, we propose an inference method of the real-life aspect distribution of tweets using labeled tweets. Our method infers the aspect probability distributions by a hierarchical estimation framework (HEF), which is hierarchically composed of both unsupervised and supervised machine learning methods. In the first phase, it extracts topics from a sea of tweets using Latent Dirichlet Allocation (LDA). In the second phase, it builds associations between topics and real-life aspects using a small set of labeled tweets. The probability distribution of aspects is inferred using the associations based on the bag of terms extracted from unknown tweets. Our sophisticated experimental evaluations with a large amount of actual tweets demonstrate the high efficiency and robustness of our inference method. Especially in the case of single-label training, HEF showed significantly lower JSD values than other baseline methods, such as Naive Bayes, SVM, and L-LDA.

Twitter reflects events and trends in users' real lives because many of them post tweets related to their experiences. Many studies have succeeded in detecting events such as earthquakes and influenza epidemics, along with real-life information from a large amount of tweets, by assuming users as social sensors. On the other hand, inactive users who don't engage in posting activity, are increasing according as time progresses. To collect a large amount of tweets based on specific users for successful Twitter studies, we have to know the characteristics of users who are active over long periods of time. In this paper, we clarify the characteristics of growth users over a long time to strategically collect a large amount of specific users' tweets. We explore the status of users who were active in 2012, and classify users into three statuses of Dead, Lock, and Alive. Based on the differences between the numbers of tweets in 2012 and 2016, we further classify alive users into three types of Eraser, Slumber, and Growth. We analyze the characteristic feature values observed in each user behavior and provide interesting findings with each status/type based on GMM clustering and point-wise mutual information. Finally, we propose a growth user prediction method by a simple formula consisting of feature values and evaluate the effectiveness. We found that active users more easily dropped out than inactive users, and users who engaged in reciprocal communications by replies and retweets often became Growth type.

Research on temporal aspects of information retrieval has recently gained considerable interest within the Information Retrieval (IR) community. This paper describes our efforts for building test collections for the purpose of fostering temporal IR research. In particular, we overview the test collections created at the two recent editions of Temporal Information Access (Temporalia) task organized at NTCIR-11 and NTCIR-12, report on selected results and discuss several observations we made during the task design and implementation. Finally, we outline further directions for constructing test collections suitable for temporal IR.

Many Twitter users build various relationships through communication activity such as replies and retweets. For example, friends engage in conversations through replies. Fans unilaterally send many replies to celebrities. In this paper, we focus on such relationships between users. We assume that such relationships are classified into several patterns based on the feature values of communication reciprocity, and the relationships migrate to other ones as time progresses. We clarify the major relationships and transitions by analyzing the pattern frequency and transitions with high probability. From analysis results using a large amount of user pairs that we obtained over a long period, we detected several major and calm relationships.

Many Twitter users post tweets that are related to their particular interests. Users can also collect information by following other users. One approach clarifies user interests by tagging labels based on the users. A user tagging method is important to discover candidate users with similar interests. Typical approaches estimate user interests with terms in tweets and by applying graph theory such as following networks. In contrast, we propose a new user tagging method using the posting time series data of the number of tweets and developed the following hypothesis: Since users have interests, they will post more tweets at the time occurring the events compared with general times. Based on this hypothesis, we extract interests as burst levels from the user and hashtag time series data with Kleinberg's burst enumerating algorithm. We manage the burst levels of users as the term frequency in documents and calculate the hashtag scores for each user by three typical score calculation methods: cosine similarity, Naive Bayes, and TF-IDF. Thus, the proposed method needs no linguistic analysis which requires heavy computational resources. With our sophisticated experimental evaluations with actually active users, we demonstrate the high efficiency of our tagging methods, evaluate them using such information retrieval system evaluation metrics as expected reciprocal rank (ERR) and Q-measure, and clarify the strengths and limitations of each one. Naive Bayes and cosine similarity are especially suitable for user tagging and tag score calculation tasks.

Recently, many users share their daily events and opinions on Twitter. Some are beneficial and comment on several aspects of a user's real life, i.e., eating, traffic, weather, disasters, and so on. Such posts as "The train is not coming!" are categorized in the "Traffic" aspect and will support users who want to ride the train. Such tweets as "The train is not coming due to heavy rain" are categorized in both the "Traffic" and "Weather" aspects. In this paper, we propose a multi-label method that estimates appropriate aspects against unknown tweets by extending the two phase extraction method. In it, many topics are extracted from a sea of tweets using Latent Dirichlet Allocation (LDA). Associations among many topics and fewer aspects are built using a small set of labeled tweets. Aspect scores for unknown tweets are calculated using the associations among the topics and the aspects based on the extracted terms. Appropriate aspects are labeled for unknown tweets by averaging of the aspect scores. Using a large amount of actual tweets, our sophisticated experimental evaluations demonstrate the high efficiency of our proposed multi-label classification method. When an aspect score is much larger than others, that aspect is estimated against the tweet. When several aspect scores are large within similar values, these aspects are estimated. Based on the experimental evaluation results, our prototype system demonstrates that our proposed method can appropriately estimate some aspects of each unknown tweets.

In recent years, such microblogs as Twitter have spread widely over the world. Twitter, which enables instant text communications among users, was launched in 2006. In 2012, its Japanese users exceeded 29.9 million. Useful functions related to posting a tweet include reply, retweet, and hashtag. Users communicate with others and spread information with these functions. In this paper, we model user behaviors by a transition of clusters that represent particular posting activities. Under the model, all users belong to a cluster consisting of several features at individual time slots and move among the clusters in a time series. These features include the number of posts and retweets/replies, the time when the tweets were posted, and the number of characters in each tweet. We reveal the temporal transitions of these clusters in the process of using Twitter from the time when users created their accounts. We propose a time longitudinal analysis method to clarify the relation of the transition of user posting activities and their lifetime of Twitter. Our proposed method consists of two steps: creating clusters that represent particular posting activities and drawing a state transition diagram with transition probabilities among clusters. From our analysis results of actual Japanese tweets for a one-year period with our proposed method, we conclude the following. Our proposed method can express changes in the posting activities of users. We conclude that using Twitter's functions, e.g., replies and retweets (RT), are one difference between users who continue to use Twitter for a long time and those who quit relatively soon.