Recently, recommendation based on causal inference has gained much attention in the industrial community. The introduction of causal techniques into recommender systems (RS) has brought great development to this field and has gradually become a trend. However, a unified causal analysis framework has not been established yet. On one hand, the existing causal methods in RS lack a clear causal and mathematical formalization on the scientific questions of interest. Many confusions need to be clarified: what exactly is being estimated, for what purpose, in which scenario, by which technique, and under what plausible assumptions. On the other hand, technically speaking, the existence of various biases is the main obstacle to drawing causal conclusions from observed data. Yet, formal definitions of the biases in RS are still not clear. Both of the limitations greatly hinder the development of RS.In this paper, we attempt to give a causal analysis framework to accommodate different scenarios in RS, thereby providing a principled and rigorous operational guideline for causal recommendation. We first propose a step-by-step guideline on how to clarify and investigate problems in RS using causal concepts. Then, we provide a new taxonomy and give a formal definition of various biases in RS from the perspective of violating what assumptions are adopted in standard causal analysis. Finally, we find that many problems in RS can be well formalized into a few scenarios using the proposed causal analysis framework.
We evaluated the effectiveness of new indices of text comprehension in measuring relative text difficulty. Specifically, we examined the efficacy of automated indices produced by the web-based computational tool Coh-Metrix. In an analysis of 60 instructional science texts, we divided texts into groups that were considered to be more or less difficult to comprehend. The defining criteria were based on Coh-Metrix indices that measure independent factors underlying text coherence: referential overlap and vocabulary accessibility. In order to validate the text difficulty groups, participants read and recalled two “difficult” and two “easy” texts that were similar in topic and length. Easier texts facilitated faster reading times and better recall compared to difficult texts. We discuss the implications of these results in the context of theoretically motivated techniques for improving textbook selection.
The following study employs RF waveform engineering to monitor degradation in 22nm FDSOI transistor at high-frequency region. The current and voltage waveforms are measured, reconstructed, and de-embedded to the device’s intrinsic during large-signal CW RF stress testing. This technique provides extra information on device performance compared with standard DC and RF figures of Merits degradation. With clear pictures of where on the output IV plane the degradation is occurring, device designers can get an insight into the degradation behavior limiting RF performance. It is observed that devices show a different behavior under RF stress in comparison to DC-stress-induced degradation.
This paper presents a highly linear 79 GHz differential low-noise amplifier (LNA) for civil-automotive radars operating at the predefined frequency range from 77 GHz to 81 GHz. The circuit is optimized for frequency-modulated continuous-wave (FMCW) radar application, which typically require a very high input-referred 1 dB-compression point (iP 1dB ). A reconfigurable differential common-source stage with capacitive neutralization is employed together with a common-gate stage in cascode configuration as the core of the LNA. The performance of the circuit can be easily adjusted within the gain-NF-P 1dB trade-off boundaries by changing the voltage at the back-gate terminal of the common-source stage, thus tailored to the application specific requirements. Passive baluns are placed at input and output to characterize the differential circuit with the available single-ended laboratory instrumentation. The LNA is implemented in a 22 nm FD-SOI CMOS technology. Its core is very compact with an area of 0.04 mm 2 . The fabricated chip is experimentally characterized in the lab, and it shows a peak gain of 8.7 dB at 80 GHz. From 75 GHz to 85 GHz, the measured input referred P 1dB (iP 1dB ) is about -6 dBm, and the minimum noise figure (NF) is 5 dB. Compared with the state-of-the-art for LNAs operating in a similar frequency range, the presented circuit shows the highest iP 1dB and has the most compact circuit core, together with an excellent NF and a moderate gain, resulting in the best figure-of-merit.
This paper presents a reflectometer with an integrated transducer as a high-integration miniaturized sensor for dielectric spectroscopy at 240 GHz in SiGe BiCMOS technology. The reflectometer consists of a signal generation component using 240-GHz multiplier chains, side-coupled directive couplers and a two-channel heterodyne receiver. Readout of the transducer upon exposure to liquids is performed by the measurement of its reflected signal using an external excitation source. The experimental dielectric sensing is demonstrated by using a binary methanol-ethanol mixture placed on the proposed on-chip dielectric sensor in the assembled printed circuit board.
The following study discusses the impact of hot-carrier degradation on high frequency performance of the 22nm FDSOI n-channel transistors. A quasi-static small-signal equivalent circuit MOSFET model is used to describe the device behavior. RF characteristics are extracted after stressing device with continuous DC. DC characteristics are also investigated thoroughly before and after stress. It is observed that, the device suffers from both interface damage and oxide defect. Accordingly, this study addresses how severe hot-carrier degradation affects the intrinsic parameters as well as the device performance.
Online community and groups often experience heated discussion. This paper examines a WeChat group discussion from the perspective of majority and minority influence to explore the evolvement of the discussion and the be-haviors of group members. Content analysis of 515 messages suggests that opin- ion conflicts between majority and minority evoke discussion engagement and knowledge exchange. There are different patterns of knowledge construction expressions between majority and minority groups. The majority prefer egocentric expression, while the minority prefer allocentric expression. Majority opinion holders have different conflict handling styles compared to minority opinion holders, who are more likely to avoid. Minority group is under great pressure in social interaction, they are easier to receive unfair comments and personal attacks.
