Figure 1.
Measured spectral irradiances of the different UV lamps and SMG absorption.
The COVID-19 pandemic has accelerated innovations for supporting learning and teaching online. However, online learning also means a reduction of opportunities in direct communication between teachers and students. Given the inevitable diversity in learning progress and achievements for individual online learners, it is difficult for teachers to give personalized guidance to a large number of students. The personalized guidance may cover many aspects, including recommending tailored exercises to a specific student according to the student′s knowledge gaps on a subject. In this paper, we propose a personalized exercise recommendation method named causal deep learning (CDL) based on the combination of causal inference and deep learning. Deep learning is used to train and generate initial feature representations for the students and the exercises, and intervention algorithms based on causal inference are then applied to further tune these feature representations. Afterwards, deep learning is again used to predict individual students′ score ratings on exercises, from which the Top-N ranked exercises are recommended to similar students who likely need enhancing of skills and understanding of the subject areas indicated by the chosen exercises. Experiments of CDL and four baseline methods on two real-world datasets demonstrate that CDL is superior to the existing methods in terms of capturing students′ knowledge gaps in learning and more accurately recommending appropriate exercises to individual students to help bridge their knowledge gaps.
Citation: Suhua Wang, Zhiqiang Ma, Hongjie Ji, Tong Liu, Anqi Chen, Dawei Zhao. Personalized exercise recommendation method based on causal deep learning: Experiments and implications[J]. STEM Education, 2022, 2(2): 157-172. doi: 10.3934/steme.2022011
| [1] | Laura Weyersberg, Florian Sommerfeld, Petra Vatter, Martin Hessling . UV radiation sensitivity of bacteriophage PhiX174 - A potential surrogate for SARS-CoV-2 in terms of radiation inactivation. AIMS Microbiology, 2023, 9(3): 431-443. doi: 10.3934/microbiol.2023023 |
| [2] | Anna-Maria Gierke, Petra Vatter, Martin Hessling . Fungal photoinactivation doses for UV radiation and visible light–a data collection. AIMS Microbiology, 2024, 10(3): 694-722. doi: 10.3934/microbiol.2024032 |
| [3] | Mushtaq Hussain, Nusrat Jabeen, Anusha Amanullah, Ayesha Ashraf Baig, Basma Aziz, Sanya Shabbir, Fozia Raza, Nasir Uddin . Molecular docking between human TMPRSS2 and SARS-CoV-2 spike protein: conformation and intermolecular interactions. AIMS Microbiology, 2020, 6(3): 350-360. doi: 10.3934/microbiol.2020021 |
| [4] | Rakhi Harne, Brittany Williams, Hazem F. M. Abdelaal, Susan L. Baldwin, Rhea N. Coler . SARS-CoV-2 infection and immune responses. AIMS Microbiology, 2023, 9(2): 245-276. doi: 10.3934/microbiol.2023015 |
| [5] | Maria I. Zapata-Cardona, Lizdany Florez-Alvarez, Ariadna L. Guerra-Sandoval, Mateo Chvatal-Medina, Carlos M. Guerra-Almonacid, Jaime Hincapie-Garcia, Juan C. Hernandez, Maria T. Rugeles, Wildeman Zapata-Builes . In vitro and in silico evaluation of antiretrovirals against SARS-CoV-2: A drug repurposing approach. AIMS Microbiology, 2023, 9(1): 20-40. doi: 10.3934/microbiol.2023002 |
| [6] | Alrayan Abass Albaz, Misbahuddin M Rafeeq, Ziaullah M Sain, Wael Abdullah Almutairi, Ali Saeed Alamri, Ahmed Hamdan Aloufi, Waleed Hassan Almalki, Mohammed Tarique . Nanotechnology-based approaches in the fight against SARS-CoV-2. AIMS Microbiology, 2021, 7(4): 368-398. doi: 10.3934/microbiol.2021023 |
| [7] | Sirio Fiorino, Andrea Carusi, Wandong Hong, Paolo Cernuschi, Claudio Giuseppe Gallo, Emanuele Ferrara, Thais Maloberti, Michela Visani, Federico Lari, Dario de Biase, Maddalena Zippi . SARS-CoV-2 vaccines: What we know, what we can do to improve them and what we could learn from other well-known viruses. AIMS Microbiology, 2022, 8(4): 422-453. doi: 10.3934/microbiol.2022029 |
| [8] | Tatyana S. Zaporozhets, Nataliya N. Besednova . Biologically active compounds from marine organisms in the strategies for combating coronaviruses. AIMS Microbiology, 2020, 6(4): 470-494. doi: 10.3934/microbiol.2020028 |
| [9] | Lucia Spicuzza, Davide Campagna, Chiara Di Maria, Enrico Sciacca, Salvatore Mancuso, Carlo Vancheri, Gianluca Sambataro . An update on lateral flow immunoassay for the rapid detection of SARS-CoV-2 antibodies. AIMS Microbiology, 2023, 9(2): 375-401. doi: 10.3934/microbiol.2023020 |
| [10] | Dharmender Kumar, Lalit Batra, Mohammad Tariq Malik . Insights of Novel Coronavirus (SARS-CoV-2) disease outbreak, management and treatment. AIMS Microbiology, 2020, 6(3): 183-203. doi: 10.3934/microbiol.2020013 |
The COVID-19 pandemic has accelerated innovations for supporting learning and teaching online. However, online learning also means a reduction of opportunities in direct communication between teachers and students. Given the inevitable diversity in learning progress and achievements for individual online learners, it is difficult for teachers to give personalized guidance to a large number of students. The personalized guidance may cover many aspects, including recommending tailored exercises to a specific student according to the student′s knowledge gaps on a subject. In this paper, we propose a personalized exercise recommendation method named causal deep learning (CDL) based on the combination of causal inference and deep learning. Deep learning is used to train and generate initial feature representations for the students and the exercises, and intervention algorithms based on causal inference are then applied to further tune these feature representations. Afterwards, deep learning is again used to predict individual students′ score ratings on exercises, from which the Top-N ranked exercises are recommended to similar students who likely need enhancing of skills and understanding of the subject areas indicated by the chosen exercises. Experiments of CDL and four baseline methods on two real-world datasets demonstrate that CDL is superior to the existing methods in terms of capturing students′ knowledge gaps in learning and more accurately recommending appropriate exercises to individual students to help bridge their knowledge gaps.
For the past two years, the world has been in the grip of the coronavirus pandemic caused by the SARS CoV-2 (coronavirus) with 500 million confirmed infections and 6.2 million recorded deaths to date [1]. SARS-CoV-2 is an enveloped single-stranded RNA virus, of zoonotic origin, which causes a respiratory disease known as COVID-19 (coronavirus disease 2019) [2],[3]. The virus is transmitted through air and is highly contagious. Coronavirus vaccines are meanwhile available, but despite approximately 10 billion doses of vaccine administered, the spread of the pandemic has not yet been halted [1]. Therefore, hygiene measures like the use of facemasks and the disinfection of surfaces or air are still mandatory. Fortunately, the coronavirus is sensitive to known chemical and physical disinfection methods, such as the application of chemical disinfectants, heat sterilization, or ultraviolet (UV) radiation [4]–[8].
However, there are processes and devices, such as air disinfection systems, where it is difficult to quantitatively assess the disinfection effect. Conducting experiments with coronaviruses requires a laboratory with biosafety level 3, which is often unavailable. Other human and animal coronaviruses such as HCoV OC43, HCoV 229E or BCoV are also pathogens but they are often employed as suitable surrogates because they are assumed to be very similar to SARS-CoV-2 and require only a biosafety level 2 laboratory [9]–[19].
Even more desirable, however, would be a SARS-CoV-2 surrogate that is nonpathogenic. Widespread bacteriophages as MS2, Q beta, PhiX174, T1, T4, T7, or Phi6 are obvious candidates, as they are not only nonpathogenic but also easy to handle. Among these phages Phi6 appears to be the most appropriate coronavirus surrogate, because like coronaviruses, Phi6 is an enveloped RNA virus, with the difference that its RNA is shorter and double-stranded [20],[21]. Additionally, in the past, the bacteriophage Phi6, has been successfully investigated or even successfully employed as a surrogate for coronaviruses in various applications [10],[22]–[32].
The study presented here will address the question of whether Phi6 is a suitable SARS-CoV-2 surrogate for virus inactivation by ultraviolet radiation. So far, three quantitative studies on the UVC (200–280 nm) sensitivity of Phi6 exist [29],[31],[33], but they differ from each other by up to a factor of 6. There is no available information of the sensitivity of Phi6 towards UVB (280–315 nm) or UVA (315–400 nm) irradiation, though these wavelengths are also known for their antimicrobial and antiviral features.
For SARS-CoV-2, there are several published UVC results available, but they differ even by a factor of about 500 between extremes, which hinders a meaningful comparison to the Phi6 properties. Additionally, some UVB and UVA SARS-CoV-2 data also exist, that could be compared to Phi6, if Phi6 results in these wavelength regions were available.
Therefore, in this study the UVC, UVB and UVA inactivation properties of Phi6 were investigated experimentally, and were compared to the results of a SARS-CoV-2 literature analysis to assess the suitability of Phi6 as a SARS-CoV-2 surrogate in potential UVC, UVB and UVA disinfection applications.
Phi6 (DSM 21518) and its host Pseudomonas syringae (DSM 21482) were obtained from Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH (Braunschweig, Germany). P. syringae was propagated in tryptic soy broth (TSB, Sigma-Aldrich, St. Louis, USA). 25 mL of TSB were inoculated with a single P. syringae colony. The culture was grown overnight at 170 rpm and 25 °C in order to obtain an optical density of 0.20 to 0.25 at 600 nm, which was equal to 1–5 x 108 colony-forming units (CFU)/mL.
A Phi6 stock solution of 109 plaque-forming units (PFU)/mL of SMG buffer (saline magnesium gelatin buffer) was prepared as described by Sambrook and Russel [34]. For the irradiation experiments, the concentrated virus was diluted in SMG by a factor of 100 to 107 PFU/mL and two 5 mL quartz beakers were filled with 2 mL each of this diluted virus sample solution. The filling height was just under 10 mm. The SMG absorption at 10 mm path length measured with a Specord Plus absorption spectrometer of Analytik Jena (Jena, Germany) is presented in Figure 1 and reveals no significant UV absorption above 240 nm.
The irradiation was performed with different ultraviolet bulbs in a distance of 32 cm above the sample: UVC (254 nm) bulb type Puritec HNS-S of Osram (Munich, Germany), UVB (311 nm) type UVB medical PLS and UVA (365 nm) PLS, both of Philips (Amsterdam, The Netherlands). The lamps were without reflectors or other optical elements. The quartz beaker containing the virus suspension was about 2 cm in diameter, which is small compared to the 32 cm distance, and thus the irradiation was assumed to be very homogeneous. The virus solution was not stirred during irradiation, however, it was mixed with the tip of the pipette just before sampling. The irradiation intensities were measured with a calibrated UV-VIS-spectroradiometer CAS 140 D of Instrument Systems (Munich, Germany).
As can be observed in Figure 1, the UV lamps also emit at longer wavelengths in the visible spectral range. However, since these emissions are weaker than the UV emissions and the antimicrobial properties of visible light are also weaker than those of UV radiation [35], these long-wavelength components were ignored.
Repeated temperature measurements with an infrared thermometer type Ranger MX4 of Raytek (Berlin, Germany) were also performed to detect a possible heating of the virus solution.
Post exposure, the Phi6 virus concentration was determined using the double agar overlay plaque assay as described by Kropinski et al. [36]: 100 µL of sequentially diluted Phi6 samples in SMG, 100 µL of host bacteria P. syringae, and 3 mL of soft TSB agar were mixed and plated over 90 mm TSB agar plates. Plaques were counted after 24 h of incubation at 25 °C and virus concentration was expressed in PFU/mL. At each sampling time point, three technical replicates were performed, and each irradiation wavelength was investigated in at least three independent experimental runs.
A literature search was performed in PubMed and Google Scholar using various combinations of the following terms: “SARS-CoV-2”, “disinfection”, “reduction”, “inactivation”, “UVC”, “UV-C”, “UVB”, “UV-B”, “UVA”, “UV-A”, “far-UVC”, “UV”, and “ultraviolet”. References in the retrieved literature were examined for their possible inclusion and references citing the identified literature were also checked.
For the analysis of the different UV sensitivities, the observed doses for a 3 log-reduction were retrieved from text, tables or figures and an average log-reduction dose (90% reduction) was calculated. Sample properties like the kind of liquid or surface were recorded. The results of SARS-CoV-2 irradiation in liquids were sorted by wavelengths and formally divided into spectral ranges of 20 nm width of 211–230 nm, 251–270 nm, 271–290 nm, 291–310 nm and 351–370 nm with the intention to get several results in each range. In the next step, median log-reduction doses were determined for each spectral range, as a high variability of the published single results-caused by different experimental setups and especially solutions with different UV absorption properties–was expected. The results of SARS-CoV-2 irradiation on surfaces were not included in the median determination, because the different surface samples of different materials and porosity are assumed to result in even larger variations as for the different liquid samples.
The measured UV irradiation intensities for UVC (200–280 nm), UVB (280–315 nm) and UVA (315–400 nm) were 0.310, 0.088 and 0.309 mW/cm2, respectively. The irradiation times were chosen so that a reduction of about 99.9 % was expected. This meant approximately 300 s, 8 h and 69 h for UVC, UVB and UVA. The temperature in the individual samples fluctuated between 18 °C and 23 °C over the entire test period, so that temperature effects could be neglected. The results of the irradiation experiments describing the change in the Phi6 concentration are presented in Figures 2, 3 and 4.
The experiments reveal a successful Phi6 photoinactivation for all three ultraviolet wavelengths. The progression of virus concentration decrease appears to be approximately exponentially dependent on irradiance, i.e., linear in the semi-logarithmic plots in Figures 2, 3, and 4. The corresponding average log-reduction doses are 31.7, 980 and 14 684 mJ/cm2 for UVC, UVB and UVA, respectively.
During the literature search, about 25 publications on UV irradiation of SARS-CoV-2 could be retrieved. Most experiments were carried out with UVC radiation at a wavelength of 254 nm, the peak emission wavelength of low-pressure mercury vapor lamps, but some authors used UV-LEDs of different wavelengths. The SARS-CoV-2 samples were in UV absorbing and non-absorbing liquids or on various surfaces, resulting in different mean log-reduction doses given in Table 1. For the probably most important spectral interval [251–270 nm], which includes the peak emission wavelength of low-pressure mercury vapor lamps, the median log-reduction dose is only 1.7 mJ/cm2.
| median log-reduction dose in mJ/cm2 [spectral range] | single result wavelength in nm | single log-reduction dose in mJ/cm2 | Sample [ref.] |
| 1.15 [211–230 nm] | 222 | 0.7 | PBS (liquid) [11] |
| 222 | 1.6 | DMEM (liquid) [37] | |
| 222 | 1.2 | plastic (surface) [38] | |
| 222 | 1.2 | plastic (surface) [39] | |
| 1.7 [251–270 nm] | 254 | 1.1 | DMEM (liquid) [40] |
| 254 | 1.3 | PBS (liquid) [11] | |
| 254 | 1.5 | PBS (liquid) [41] | |
| 254 | 1.7 | DMEM/PBS [19] | |
| 254 | 2 | DMEM? (liquid) [42] | |
| 254 | 2.2 | DMEM-HD (liquid) [43] | |
| 254 | 2.7 | DMEM (liquid) [44] | |
| 254 | 5.3 | DMEM (liquid) [45] | |
| 254 | 5.3 | DMEM (liquid) [46] | |
| 254 | *reduction observed but no dose available | airway epithelial cells ("liquid") [47] | |
| 254 | *< 0.6 | plastic (surface) [48] | |
| 254 | *1.7 "wet" *1.3 "dry" |
"wet" (liquid) "dry" (plastic surface) [49] |
|
| 254 | *5.4 @ plastic *3.9 @ steel *<2.2 @ glas |
plastic, steel, glass (surface) [50] | |
| 254 | *10 @ plastic *16 @ makeup *17 @ lipstick |
plastic, powder and lipstick (surface) [51] | |
| 254 | *< 540 | plastic (surface) [52] | |
| 254 | *reduction observed | different N95 respirators [53] | |
| 254 | *reduction observed but no dose available | plastic (surface) [54] | |
| 265 | 1.5 | PBS (liquid) [41] | |
| 265 | 0.6 | different media (liquid) [55] | |
| 267 | 1.9 | PBS (liquid) [9] | |
| 270 | 1.1 | PBS (liquid) [11] | |
| 270 | *4.4 | glass (surface) [56] | |
| 2.3 [271–290 nm] | 275 | reduction observed but no dose available | unknown (liquid) [57] |
| 278 | 1 | DMEM (liquid) [58] | |
| 279 | 2.3 | PBS (liquid) [9] | |
| 280 | 2.8 | PBS (liquid) [41] | |
| 280 | 12.1 | PBS (liquid) [59] | |
| 280 | 1 | different media (liquid) [55] | |
| 282 | 1.9 | PBS (liquid) [11] | |
| 286 | 4.3 | PBS (liquid) [9] | |
| 10.7 [291–310 nm] | 297 | 10.7 | PBS (liquid) [9] |
| 300 | 7.7 | different media (liquid) [55] | |
| 308 | 154 | DMEM (liquid) [58] | |
| 2 569 [351–370 nm] | 365 | 278 | DMEM? (liquid) [42] |
| 366 | 4860 | DMEM (liquid) [58] | |
| UVC & UVB & UVA (pulsed xenon light) | *reduction observed but no dose available | plastic and N95 respirators (surface) [60] | |
| UVB & UVA (simulated sun light) | *207 (UVB) | PBS (liquid) [61] | |
| UVB & UVA (simulated sunlight) | *33.6 (UVB) @ saliva *73.7 (UVB) @ GMEM |
GMEM or simulated saliva on steel (surface) [62] | |
*Note: Results marked with “*” were not included in the determination of the median, because they were obtained on very different surfaces or information, e.g. irradiation dose was missing. (DMEM, EMEM, GMEM: different cell culture media with strong UV absorption, PBS: phosphate buffered saline with low UV absorption.)
DownLoad:
CSV
All UV spectral ranges examined in the present work show antiviral properties towards Phi6, with UVC proving to be the most effective and UVA the least one. Table 2 lists a comparison of the results presented here with the radiation doses for a Phi6 log-reduction already described in the literature. The here observed Phi6 UVC log-reduction dose of 31.7 mJ/cm2 is more than twice as high as the value reported by Ye et al. [29] but in in excellent agreement with the results of Ma et al. [31]. If this study is included in the determination of the median, the Phi6 log-reduction dose is about 31.5 mJ/cm2.
| median log-reduction dose in mJ/cm2 [spectral range] | single result wavelength in nm | single log-reduction dose in mJ/cm2 | Sample [ref.] |
| 2.8 [211–230 nm] | 222 | 2.8 | PBS (liquid) [31] |
| 24.1 [251–270 nm] 31.5** |
254 | 14.9 | PBS (liquid) [29] |
| 254 | 33.3 | PBS (liquid) [31] | |
| 254 | 31.7 | SMG (liquid) [this study] | |
| 254 | *5.3 *6.0 |
gelatin-medium (surface) [33] | |
| 254 | *reduction observed but no dose available | steel disks, N95 respirator and plastic bins (surface) [30],[63] | |
| 270 | 31.3 | PBS (liquid) [31] | |
| 40 [271–290 nm] | 282 | 40 | PBS (liquid) [31] |
| 98 [311–330 nm] | 311 | 980 | SMG (liquid) [this study] |
| 14 684 [351–370 nm] | 365 | 14 684 | SMG (liquid) [this study] |
*Note: Results marked with “*” were not included in the determination of the median, because they were obtained on very different surfaces or information like irradiation dose was missing. “**” median log-reduction dose for Phi6 including the results from this study. (PBS: phosphate buffered saline, SMG: saline magnesium gelatin buffer, both exhibit low UV absorption.)
DownLoad:
CSV
For the UVB and UVA data there are no published Phi6 log-reduction doses for comparison, but some UV results for the ssRNA bacteriophage MS2, with log-reduction doses of about 7 mJ/cm2 at 222 nm, 14–30 mJ/cm2 in the spectral interval [251–270 nm], 20–63 mJ/cm2 in the interval [271–290 nm], 230–330 mJ/cm2 in the UVB and 910 mJ/cm2 in the UVA range [64]–[67]. The UVC properties of MS2 seem to be very similar to Phi6 in the UVC spectral range. In the UVB and UVA ranges, there is also a large increase for the MS2 log-reduction dose, but–with the limited available data-it seems to be less distinctive than with Phi6, which might just be caused by different applied irradiation spectra.
Unfortunately, the available Phi6 and SARS-CoV-2 data are still sparse and exhibit a high variability. Nevertheless, at least in the spectral range [251–270 nm] data on the Phi6 and SARS-CoV-2 UV sensitivity from different authors are available that seem to fit to another. The results from the other spectral regions are less well funded, but they give a consistent overall picture as shown in Table 3.
Both viruses can be inactivated by UVC, UVB and UVA irradiation, and the here presented relative UV sensitivities decrease with longer wavelengths. However, though both viruses are enveloped RNA viruses the absolute UV sensitivity or necessary log-reduction doses differ, with Phi6 being much more UV resistant than SARS-CoV-2 for UVC, UVB and UVA. This might be caused by differences in the RNA structure and length. Though protein damage might contribute, the main virus inactivation mechanism seems to be caused by RNA/DNA damage [66],[68]–[70]. ssRNA viruses seem to be more UV sensitive than dsRNA ones and longer RNA is more susceptible to ultraviolet radiation than shorter ones [70]. Both properties lead to the result that the 13.5 kbp dsRNA virus Phi6 exhibits higher log-reduction doses than the 30 kbp ssRNA SARS-CoV-2.
| Spectral interval | Phi6 median log-reduction dose in mJ/cm2 | SARS-CoV-2 median log-reduction dose in mJ/cm2 | ratio |
| [211–230 nm] | 2.8 | 1.15 | 2.4 |
| [251–270 nm] | 31.5 | 1.7 | 18.5 |
| [271–290 nm] | 40 | 2.3 | 17.4 |
| [291–310 nm] | 10.7 | 91.6 | |
| [311–330 nm] | 980 | ||
| [351–370 nm] | 14 684 | 2 569 | 5.7 |
*Note: UVB results for Phi6 and SARS-Co-V-2 are formerly in two different spectral regions but in fact, they are less than 15 nm apart and therefore compared to another.
DownLoad:
CSV
This confirms the experimental observation that despite structural similarities Phi6 is less UV sensitive than SARS-CoV-2 and therefore Phi6 is not the best suited surrogate for SARS-CoV-2 in UVC, UVB or UVA applications.
| [1] |
Vie, J. and Kashima, H., Knowledge tracing machines: Factorization machines for knowledge tracing. The Thirty-Third AAAI Conference on Artificial Intelligence, 2019, 33(01): 750–757. https://doi.org/10.1609/aaai.v33i01.3301750 doi: 10.1609/aaai.v33i01.3301750
|
| [2] |
Walker, A., Recker, M. and Lawless, K., Collaborative information filtering: A review and an educational application. International Journal of Artificial Intelligence in Education, 2004, 14(1): 3–28. |
| [3] |
Hsu, M., A personalized English learning recommender system for ESL students. Expert Systems with Applications, 2008, 34(1): 683–688. https://doi.org/10.1016/j.eswa.2006.10.004 doi: 10.1016/j.eswa.2006.10.004
|
| [4] |
Milicevic, A., Vesin, B. and Ivanovic, M., E-learning personalization based on hybrid recommendation strategy and learning style identification. Computers & Education, 2011, 56(3): 885–899. https://doi.org/10.1016/j.compedu.2010.11.001 doi: 10.1016/j.compedu.2010.11.001
|
| [5] |
Segal, A., Katzir, Z. and Shapira, B., EduRank: A collaborative filtering approach to personalization in e-learning. Proceedings of the 7th International Conference on Educational Data Mining, 2014, 68–75. |
| [6] |
Toledo, R. and Mota, Y., An e-learning collaborative filtering approach to suggest problems to solve in programming online judges. International Journal of Distance Education Technologies, 2014, 12(2): 51–65. https://doi.org/10.4018/ijdet.2014040103 doi: 10.4018/ijdet.2014040103
|
| [7] |
Wu, D., Lu, J. and Zhang G., A fuzzy tree matching-based personalized e-learning recommender system. IEEE Transactions on Fuzzy Systems, 2015, 23(6): 2412–2426. https://doi.org/10.1109/TFUZZ.2015.2426201 doi: 10.1109/TFUZZ.2015.2426201
|
| [8] |
Dwivedi, P. and Bharadwaj, K., Effective trust-aware e-learning recommender system based on learning styles and knowledge levels. Journal of Educational Technology & Society, 2013, 16(4): 201–216. |
| [9] |
Jiang, C., Feng, J. and Sun, X., Personalized exercises recommendation algorithm based on knowledge hierarchical graph, ReKHG. Computer Engineering and Applications, 2018, 54(10): 234–240. |
| [10] |
Gong, T. and Yao, X., Deep exercise recommendation model. International Journal of Modeling and Optimization, 2019, 9(1): 18–23. https://doi.org/10.7763/IJMO.2019.V9.677 doi: 10.7763/IJMO.2019.V9.677
|
| [11] |
Piech, C., Bassen, J. and Huang, J., Deep knowledge tracing. NIPS'15 Proceedings of the 28th International Conference on Neural Information Processing Systems, 2015, 505–513. |
| [12] |
Zhang, L., Xiong, X. and Zhao, S., Incorporating rich features into deep knowledge tracing. Proceedings of the Fourth ACM Conference on Learning at Scale, 2017, 169–172. https://doi.org/10.1145/3051457.3053976 doi: 10.1145/3051457.3053976
|
| [13] |
Su, Y., Liu, Q. and Liu, Q., Exercise-enhanced sequential modeling for student performance prediction. Proceedings of the 32nd AAAI Conference on Artificial Intelligence (AAAI-18), 2018, 2435–2443. |
| [14] |
Wang, L., Angela, S. and Liu L., Deep knowledge tracing on programming exercises. Proceedings of the Fourth. ACM Conference on Learning at Scale, 2017, 201–204. https://doi.org/10.1145/3051457.3053985 doi: 10.1145/3051457.3053985
|
| [15] |
Yeung, C. and Yeung, D., Addressing two problems in deep knowledge tracing via prediction-consistent regularization. Proceedings of the Fifth Annual ACM Conference on Learning at Scale, 2018, 41–50. https://doi.org/10.1145/3231644.3231647 doi: 10.1145/3231644.3231647
|
| [16] |
Didelez, V. and Pigeot, I., Judea Pearl: Causality: Models, reasoning, and inference. Politische Vierteljahresschrift, 2001, 42(2): 313–315. https://doi.org/10.1007/s11615-001-0048-3 doi: 10.1007/s11615-001-0048-3
|
| [17] |
Louizos, C., Shalit, U. and Mooij, J., Causal effect inference with deep latent-variable models. Proceedings of the 31st International Conference on Neural Information Processing Systems, 2017, 6449–6459. |
| [18] |
Van, D., Causal reasoning and inference making in judging the importance of story statements. Child Development, 1989, 60(2): 286–297. |
| [19] |
Joachims, T., Swaminathan, A. and De, R., Deep learning with logged bandit feedback. International Conference on Learning Representations, 2018, 1–12. |
| [20] |
Swaminathan, A. and Joachims, T., Counterfactual risk minimization: Learning from logged bandit feedback. Proceedings of the 32nd International Conference on Machine Learning, 2015, 814–823. https://doi.org/10.1145/2740908.2742564 doi: 10.1145/2740908.2742564
|
| [21] |
Swaminathan, A. and Joachims, T., The self-normalized estimator for counterfactual learning. Advances in Neural Information Processing Systems, 2015, 3231–3239. |
| [22] |
Lisa, W., Angela, S., Larry. L. and Chris, P., Deep knowledge tracing on programming exercises. Proceedings of the Fourth ACM Conference on Learning, 2017, 201–204. https://doi.org/10.1145/3051457.3053985 doi: 10.1145/3051457.3053985
|
| [23] |
Guy, S. and Asela, G., Evaluating recommendation systems. Recommender Systems Handbook, 2011, 257–297. https://doi.org/10.1007/978-0-387-85820-3_8 doi: 10.1007/978-0-387-85820-3_8
|
| [24] |
Gang, L. and Tianyong, H., User-based question recommendation for question answering system. International Journal of Information and Education Technology, 2012, 2(3): 243–246. https://doi.org/10.7763/IJIET.2012.V2.120 doi: 10.7763/IJIET.2012.V2.120
|
| [25] |
Shah, K., Zafar, A. and Irfan, U., Recommender systems: Issues, challenges, and research opportunities. Information science and applications (ICISA) 2016, 2016, 1179–1189. https://doi.org/10.1007/978-981-10-0557-2_112 doi: 10.1007/978-981-10-0557-2_112
|
| [26] |
Ming, Z., De-sheng, Z., Ran, T., You-Qun, S., Xiang, Y. and Qian, W., Top-N collaborative filtering recommendation algorithm based on knowledge graph embedding. Proceedings of the 14th International Conference of the Knowledge Management in Organizations, 2019, 122–134. https://doi.org/10.1007/978-3-030-21451-7_11 doi: 10.1007/978-3-030-21451-7_11
|
| 1. | Gerardo Grasso, Daniela Zane, Sabrina Foglia, Roberto Dragone, Application of Electrospun Water-Soluble Synthetic Polymers for Multifunctional Air Filters and Face Masks, 2022, 27, 1420-3049, 8753, 10.3390/molecules27248753 | |
| 2. | Martin Hessling, Ben Sicks, Bernhard Lau, Far-UVC Radiation for Disinfecting Hands or Gloves?, 2023, 12, 2076-0817, 213, 10.3390/pathogens12020213 | |
| 3. | Trailokya Bhattarai, Abasifreke Ebong, M. Yasin Akhtar Raja, Mariya Munir, 2023, Advancements in 275 nm UV-LED Technology for Deactivation of Bacteriophages, Phi6 and MS2, 979-8-3503-3111-0, 120, 10.1109/HONET59747.2023.10374888 | |
| 4. | Florian Sommerfeld, Laura Weyersberg, Petra Vatter, Martin Hessling, Photoinactivation of the bacteriophage PhiX174 by UVA radiation and visible light in SM buffer and DMEM-F12, 2024, 17, 1756-0500, 10.1186/s13104-023-06658-8 | |
| 5. | Rasmus Malmgren, Svetlana Sofieva, Kaisa Lakkala, Eija Asmi, Yan Sun, Michael Todt, Paavo Heikkilä, Dennis Bamford, Martin Romantschuk, Nina Atanasova, Phi6 virus transmission in Arctic outdoor air and the effects of solar UV radiation on virus and host viability, 2024, 26667657, 100600, 10.1016/j.envadv.2024.100600 | |
| 6. | Xing Qiu, Jeffery C. C. Lo, Yuanjie Cheng, Hua Xu, Qianwen Xu, Shi-Wei Ricky Lee, An Efficient Contamination-Reducing Closet for Reusing Protective Clothing, 2025, 10, 2411-5134, 18, 10.3390/inventions10010018 |
Figures(5) / Tables(7)
Suhua Wang, Zhiqiang Ma, Hongjie Ji, Tong Liu, Anqi Chen, Dawei Zhao. Personalized exercise recommendation method based on causal deep learning: Experiments and implications[J]. STEM Education, 2022, 2(2): 157-172. doi: 10.3934/steme.2022011
| median log-reduction dose in mJ/cm2 [spectral range] | single result wavelength in nm | single log-reduction dose in mJ/cm2 | Sample [ref.] |
| 1.15 [211–230 nm] | 222 | 0.7 | PBS (liquid) [11] |
| 222 | 1.6 | DMEM (liquid) [37] | |
| 222 | 1.2 | plastic (surface) [38] | |
| 222 | 1.2 | plastic (surface) [39] | |
| 1.7 [251–270 nm] | 254 | 1.1 | DMEM (liquid) [40] |
| 254 | 1.3 | PBS (liquid) [11] | |
| 254 | 1.5 | PBS (liquid) [41] | |
| 254 | 1.7 | DMEM/PBS [19] | |
| 254 | 2 | DMEM? (liquid) [42] | |
| 254 | 2.2 | DMEM-HD (liquid) [43] | |
| 254 | 2.7 | DMEM (liquid) [44] | |
| 254 | 5.3 | DMEM (liquid) [45] | |
| 254 | 5.3 | DMEM (liquid) [46] | |
| 254 | *reduction observed but no dose available | airway epithelial cells ("liquid") [47] | |
| 254 | *< 0.6 | plastic (surface) [48] | |
| 254 | *1.7 "wet" *1.3 "dry" |
"wet" (liquid) "dry" (plastic surface) [49] |
|
| 254 | *5.4 @ plastic *3.9 @ steel *<2.2 @ glas |
plastic, steel, glass (surface) [50] | |
| 254 | *10 @ plastic *16 @ makeup *17 @ lipstick |
plastic, powder and lipstick (surface) [51] | |
| 254 | *< 540 | plastic (surface) [52] | |
| 254 | *reduction observed | different N95 respirators [53] | |
| 254 | *reduction observed but no dose available | plastic (surface) [54] | |
| 265 | 1.5 | PBS (liquid) [41] | |
| 265 | 0.6 | different media (liquid) [55] | |
| 267 | 1.9 | PBS (liquid) [9] | |
| 270 | 1.1 | PBS (liquid) [11] | |
| 270 | *4.4 | glass (surface) [56] | |
| 2.3 [271–290 nm] | 275 | reduction observed but no dose available | unknown (liquid) [57] |
| 278 | 1 | DMEM (liquid) [58] | |
| 279 | 2.3 | PBS (liquid) [9] | |
| 280 | 2.8 | PBS (liquid) [41] | |
| 280 | 12.1 | PBS (liquid) [59] | |
| 280 | 1 | different media (liquid) [55] | |
| 282 | 1.9 | PBS (liquid) [11] | |
| 286 | 4.3 | PBS (liquid) [9] | |
| 10.7 [291–310 nm] | 297 | 10.7 | PBS (liquid) [9] |
| 300 | 7.7 | different media (liquid) [55] | |
| 308 | 154 | DMEM (liquid) [58] | |
| 2 569 [351–370 nm] | 365 | 278 | DMEM? (liquid) [42] |
| 366 | 4860 | DMEM (liquid) [58] | |
| UVC & UVB & UVA (pulsed xenon light) | *reduction observed but no dose available | plastic and N95 respirators (surface) [60] | |
| UVB & UVA (simulated sun light) | *207 (UVB) | PBS (liquid) [61] | |
| UVB & UVA (simulated sunlight) | *33.6 (UVB) @ saliva *73.7 (UVB) @ GMEM |
GMEM or simulated saliva on steel (surface) [62] | |
*Note: Results marked with “*” were not included in the determination of the median, because they were obtained on very different surfaces or information, e.g. irradiation dose was missing. (DMEM, EMEM, GMEM: different cell culture media with strong UV absorption, PBS: phosphate buffered saline with low UV absorption.)
DownLoad:
CSV
| median log-reduction dose in mJ/cm2 [spectral range] | single result wavelength in nm | single log-reduction dose in mJ/cm2 | Sample [ref.] |
| 2.8 [211–230 nm] | 222 | 2.8 | PBS (liquid) [31] |
| 24.1 [251–270 nm] 31.5** |
254 | 14.9 | PBS (liquid) [29] |
| 254 | 33.3 | PBS (liquid) [31] | |
| 254 | 31.7 | SMG (liquid) [this study] | |
| 254 | *5.3 *6.0 |
gelatin-medium (surface) [33] | |
| 254 | *reduction observed but no dose available | steel disks, N95 respirator and plastic bins (surface) [30],[63] | |
| 270 | 31.3 | PBS (liquid) [31] | |
| 40 [271–290 nm] | 282 | 40 | PBS (liquid) [31] |
| 98 [311–330 nm] | 311 | 980 | SMG (liquid) [this study] |
| 14 684 [351–370 nm] | 365 | 14 684 | SMG (liquid) [this study] |
*Note: Results marked with “*” were not included in the determination of the median, because they were obtained on very different surfaces or information like irradiation dose was missing. “**” median log-reduction dose for Phi6 including the results from this study. (PBS: phosphate buffered saline, SMG: saline magnesium gelatin buffer, both exhibit low UV absorption.)
DownLoad:
CSV
| Spectral interval | Phi6 median log-reduction dose in mJ/cm2 | SARS-CoV-2 median log-reduction dose in mJ/cm2 | ratio |
| [211–230 nm] | 2.8 | 1.15 | 2.4 |
| [251–270 nm] | 31.5 | 1.7 | 18.5 |
| [271–290 nm] | 40 | 2.3 | 17.4 |
| [291–310 nm] | 10.7 | 91.6 | |
| [311–330 nm] | 980 | ||
| [351–370 nm] | 14 684 | 2 569 | 5.7 |
*Note: UVB results for Phi6 and SARS-Co-V-2 are formerly in two different spectral regions but in fact, they are less than 15 nm apart and therefore compared to another.
DownLoad:
CSV
| median log-reduction dose in mJ/cm2 [spectral range] | single result wavelength in nm | single log-reduction dose in mJ/cm2 | Sample [ref.] |
| 1.15 [211–230 nm] | 222 | 0.7 | PBS (liquid) [11] |
| 222 | 1.6 | DMEM (liquid) [37] | |
| 222 | 1.2 | plastic (surface) [38] | |
| 222 | 1.2 | plastic (surface) [39] | |
| 1.7 [251–270 nm] | 254 | 1.1 | DMEM (liquid) [40] |
| 254 | 1.3 | PBS (liquid) [11] | |
| 254 | 1.5 | PBS (liquid) [41] | |
| 254 | 1.7 | DMEM/PBS [19] | |
| 254 | 2 | DMEM? (liquid) [42] | |
| 254 | 2.2 | DMEM-HD (liquid) [43] | |
| 254 | 2.7 | DMEM (liquid) [44] | |
| 254 | 5.3 | DMEM (liquid) [45] | |
| 254 | 5.3 | DMEM (liquid) [46] | |
| 254 | *reduction observed but no dose available | airway epithelial cells ("liquid") [47] | |
| 254 | *< 0.6 | plastic (surface) [48] | |
| 254 | *1.7 "wet" *1.3 "dry" |
"wet" (liquid) "dry" (plastic surface) [49] |
|
| 254 | *5.4 @ plastic *3.9 @ steel *<2.2 @ glas |
plastic, steel, glass (surface) [50] | |
| 254 | *10 @ plastic *16 @ makeup *17 @ lipstick |
plastic, powder and lipstick (surface) [51] | |
| 254 | *< 540 | plastic (surface) [52] | |
| 254 | *reduction observed | different N95 respirators [53] | |
| 254 | *reduction observed but no dose available | plastic (surface) [54] | |
| 265 | 1.5 | PBS (liquid) [41] | |
| 265 | 0.6 | different media (liquid) [55] | |
| 267 | 1.9 | PBS (liquid) [9] | |
| 270 | 1.1 | PBS (liquid) [11] | |
| 270 | *4.4 | glass (surface) [56] | |
| 2.3 [271–290 nm] | 275 | reduction observed but no dose available | unknown (liquid) [57] |
| 278 | 1 | DMEM (liquid) [58] | |
| 279 | 2.3 | PBS (liquid) [9] | |
| 280 | 2.8 | PBS (liquid) [41] | |
| 280 | 12.1 | PBS (liquid) [59] | |
| 280 | 1 | different media (liquid) [55] | |
| 282 | 1.9 | PBS (liquid) [11] | |
| 286 | 4.3 | PBS (liquid) [9] | |
| 10.7 [291–310 nm] | 297 | 10.7 | PBS (liquid) [9] |
| 300 | 7.7 | different media (liquid) [55] | |
| 308 | 154 | DMEM (liquid) [58] | |
| 2 569 [351–370 nm] | 365 | 278 | DMEM? (liquid) [42] |
| 366 | 4860 | DMEM (liquid) [58] | |
| UVC & UVB & UVA (pulsed xenon light) | *reduction observed but no dose available | plastic and N95 respirators (surface) [60] | |
| UVB & UVA (simulated sun light) | *207 (UVB) | PBS (liquid) [61] | |
| UVB & UVA (simulated sunlight) | *33.6 (UVB) @ saliva *73.7 (UVB) @ GMEM |
GMEM or simulated saliva on steel (surface) [62] | |
| median log-reduction dose in mJ/cm2 [spectral range] | single result wavelength in nm | single log-reduction dose in mJ/cm2 | Sample [ref.] |
| 2.8 [211–230 nm] | 222 | 2.8 | PBS (liquid) [31] |
| 24.1 [251–270 nm] 31.5** |
254 | 14.9 | PBS (liquid) [29] |
| 254 | 33.3 | PBS (liquid) [31] | |
| 254 | 31.7 | SMG (liquid) [this study] | |
| 254 | *5.3 *6.0 |
gelatin-medium (surface) [33] | |
| 254 | *reduction observed but no dose available | steel disks, N95 respirator and plastic bins (surface) [30],[63] | |
| 270 | 31.3 | PBS (liquid) [31] | |
| 40 [271–290 nm] | 282 | 40 | PBS (liquid) [31] |
| 98 [311–330 nm] | 311 | 980 | SMG (liquid) [this study] |
| 14 684 [351–370 nm] | 365 | 14 684 | SMG (liquid) [this study] |
| Spectral interval | Phi6 median log-reduction dose in mJ/cm2 | SARS-CoV-2 median log-reduction dose in mJ/cm2 | ratio |
| [211–230 nm] | 2.8 | 1.15 | 2.4 |
| [251–270 nm] | 31.5 | 1.7 | 18.5 |
| [271–290 nm] | 40 | 2.3 | 17.4 |
| [291–310 nm] | 10.7 | 91.6 | |
| [311–330 nm] | 980 | ||
| [351–370 nm] | 14 684 | 2 569 | 5.7 |
Framework for causal deep learning (CDL) (Cs is the student input embedded with causal interventions; Cs is the exercise input embedded with causal interventions)
Influence of the length of the knowledge path
RMSE with different embedding sizes
RMSE with different epochs
Impact of interaction layers
