PMC:7795931 / 26835-32184 JSON TXT

Annnotations TAB JSON ListView MergeView

LitCovid-PubTator

LitCovid-PD-HP

{"project":"LitCovid-PD-HP","denotations":[{"id":"T3","span":{"begin":1027,"end":1032},"obj":"Phenotype"},{"id":"T4","span":{"begin":1267,"end":1272},"obj":"Phenotype"},{"id":"T5","span":{"begin":2460,"end":2465},"obj":"Phenotype"}],"attributes":[{"id":"A3","pred":"hp_id","subj":"T3","obj":"http://purl.obolibrary.org/obo/HP_0002527"},{"id":"A4","pred":"hp_id","subj":"T4","obj":"http://purl.obolibrary.org/obo/HP_0002527"},{"id":"A5","pred":"hp_id","subj":"T5","obj":"http://purl.obolibrary.org/obo/HP_0002527"}],"text":"3.2. Intervention Dilemma in Disclosing Information\nIn this part, we will discuss the speed-accuracy trade-off results and analyze the mechanism in information disclosing.\nFirst, Figure 3 summarizes the results of the simulations with 44 different external constraints, and we find that there is seldom a single dominant disclosing threshold (the government’s preference on speed or accuracy), i.e., seeking either speed or accuracy alone will not result in the lowest infection rate, and the optimal strategy (corresponding to the lowest infection rate at the end of the last period) is somewhere in between, which implies a speed-accuracy trade-off.\nSpecifically, the optimal disclosing threshold lies between both ends in about 84.09% of the cases, and their distributions vary in different external constraints. Figure 3a,b show that (1) if the initial information is 1 or 0.8 and the individual threshold is in [0,0.7], the optimal disclosing threshold has a 91.75% probability of being in the middle, and mostly (66.89%) falls within [0.6,0.8]; (2) if the initial information is 0.6 and the individual threshold is in [0,0.5], the government has a 93.00% probability of dealing with a trade-off (Figure 3c), and the optimal disclosing threshold mostly (77.78%) falls in [0.4,0.6]; (3) if the initial information is 0.4, the optimal disclosing threshold will almost certainly be greater than 0.4 (99.98%), but the distribution is too scattered to give a specific interval (Figure 3d). The mode of optimal disclosing threshold is 0.8 but only with 20.30% frequency.\nIn addition, we can see that, if the virus is medically well-known (ξ≥0.8) and public health awareness is low (XI≥0.7), the government shall prioritize accuracy over speed; if the virus is medically medium-known (ξ=0.6) and public awareness is high (XI≤0.5), the government shall balance speed and accuracy, which almost equally signifies; and if the virus in medically less-known (ξ=0.4), the government shall probably prioritize accuracy over speed.\nSecond, we will dissect the underlying logic and mechanism of the government’s trade-off. We take one of the curves in Figure 3a that is denoted by ξ=1 and XI=0.5, as an example, to find the relationship between disclosing the threshold infection rate, then we have Figure 4. In all 550 (11×50) experiments, the disclosing threshold for the lowest final infection rate usually lies between 0.7 and 0.9, and the final infection rate first falls then rises as the disclosing threshold increases, with the inflection point being at 0.8 (Figure 4a); the amount of final information per capita and the duration of government intervention both increase monotonically with the disclosing threshold (Figure 4b,e); the number of people infected after government intervention, the number of people infected by a panic after government intervention and the number of uninfected people remaining at the time of government intervention all negatively correlate with the disclosing threshold (Figure 4c,d,f).\nOne of the fundamental reasons for the government to balance speed and accuracy is the precipitous fall in the marginal contribution of accuracy as the disclosing threshold exceeds a certain “point”, while speed hardly affects the final infection rate. The following is a detailed analysis on the effects of both accuracy and speed.\nWith respect to accuracy, the effect comes from two perspectives: (1) accurate information lowers the infection from panic in a healthy panic-prone population (Figure 4d); (2) accurate information changes the behavior routine in the panic-prone population, which reduces the spread of the disease. Both (1) and (2) are in play until the disclosing threshold exceeds 0.5 (XI in our example); after that, there is no longer a panic-prone group, neither is infection from panic, which explains the precipitous fall in the marginal contribution of accuracy.\nWhen it comes to speed, the effect comes from two perspectives as well: (1) the time the government spends on censoring and screening information, which we call a lag; (2) the number of remaining uninfected people at the time of government disclosing information. The more accurate information the government seeks, the longer the lag (Figure 4e) and the fewer uninfected people remain at the time of disclosing (Figure 4f). Notice that both have roughly the same slope with respect to disclosing threshold, which explains a roughly constant marginal cost of pursuing accuracy.\nThe combination of constant marginal costs and abrupt fall in marginal benefits leads to an inflection point in disclosing threshold, which explains the heterogeneity in the distribution of optimal disclosing threshold: as disclosing threshold exceeds individual threshold, a sudden fall of benefits occurs, which theoretically makes the optimal disclosing threshold slightly greater than the individual threshold, which explains what we discussed above that the intervals in which the optimal disclosing threshold mostly lays differ.\nIn most cases, the government’s premature disclosing of inaccurate information will contaminate the overall network, while obsession with accuracy may have the government miss the disclosure window before too many people are infected, which is intolerable to government who requires a low infection rate. Thus, there is a trade-off."}

LitCovid-sentences

{"project":"LitCovid-sentences","denotations":[{"id":"T179","span":{"begin":0,"end":4},"obj":"Sentence"},{"id":"T180","span":{"begin":5,"end":51},"obj":"Sentence"},{"id":"T181","span":{"begin":52,"end":171},"obj":"Sentence"},{"id":"T182","span":{"begin":172,"end":651},"obj":"Sentence"},{"id":"T183","span":{"begin":652,"end":815},"obj":"Sentence"},{"id":"T184","span":{"begin":816,"end":1489},"obj":"Sentence"},{"id":"T185","span":{"begin":1490,"end":1569},"obj":"Sentence"},{"id":"T186","span":{"begin":1570,"end":2021},"obj":"Sentence"},{"id":"T187","span":{"begin":2022,"end":2111},"obj":"Sentence"},{"id":"T188","span":{"begin":2112,"end":2297},"obj":"Sentence"},{"id":"T189","span":{"begin":2298,"end":3016},"obj":"Sentence"},{"id":"T190","span":{"begin":3017,"end":3269},"obj":"Sentence"},{"id":"T191","span":{"begin":3270,"end":3349},"obj":"Sentence"},{"id":"T192","span":{"begin":3350,"end":3647},"obj":"Sentence"},{"id":"T193","span":{"begin":3648,"end":3903},"obj":"Sentence"},{"id":"T194","span":{"begin":3904,"end":4167},"obj":"Sentence"},{"id":"T195","span":{"begin":4168,"end":4328},"obj":"Sentence"},{"id":"T196","span":{"begin":4329,"end":4481},"obj":"Sentence"},{"id":"T197","span":{"begin":4482,"end":5016},"obj":"Sentence"},{"id":"T198","span":{"begin":5017,"end":5321},"obj":"Sentence"},{"id":"T199","span":{"begin":5322,"end":5349},"obj":"Sentence"}],"namespaces":[{"prefix":"_base","uri":"http://pubannotation.org/ontology/tao.owl#"}],"text":"3.2. Intervention Dilemma in Disclosing Information\nIn this part, we will discuss the speed-accuracy trade-off results and analyze the mechanism in information disclosing.\nFirst, Figure 3 summarizes the results of the simulations with 44 different external constraints, and we find that there is seldom a single dominant disclosing threshold (the government’s preference on speed or accuracy), i.e., seeking either speed or accuracy alone will not result in the lowest infection rate, and the optimal strategy (corresponding to the lowest infection rate at the end of the last period) is somewhere in between, which implies a speed-accuracy trade-off.\nSpecifically, the optimal disclosing threshold lies between both ends in about 84.09% of the cases, and their distributions vary in different external constraints. Figure 3a,b show that (1) if the initial information is 1 or 0.8 and the individual threshold is in [0,0.7], the optimal disclosing threshold has a 91.75% probability of being in the middle, and mostly (66.89%) falls within [0.6,0.8]; (2) if the initial information is 0.6 and the individual threshold is in [0,0.5], the government has a 93.00% probability of dealing with a trade-off (Figure 3c), and the optimal disclosing threshold mostly (77.78%) falls in [0.4,0.6]; (3) if the initial information is 0.4, the optimal disclosing threshold will almost certainly be greater than 0.4 (99.98%), but the distribution is too scattered to give a specific interval (Figure 3d). The mode of optimal disclosing threshold is 0.8 but only with 20.30% frequency.\nIn addition, we can see that, if the virus is medically well-known (ξ≥0.8) and public health awareness is low (XI≥0.7), the government shall prioritize accuracy over speed; if the virus is medically medium-known (ξ=0.6) and public awareness is high (XI≤0.5), the government shall balance speed and accuracy, which almost equally signifies; and if the virus in medically less-known (ξ=0.4), the government shall probably prioritize accuracy over speed.\nSecond, we will dissect the underlying logic and mechanism of the government’s trade-off. We take one of the curves in Figure 3a that is denoted by ξ=1 and XI=0.5, as an example, to find the relationship between disclosing the threshold infection rate, then we have Figure 4. In all 550 (11×50) experiments, the disclosing threshold for the lowest final infection rate usually lies between 0.7 and 0.9, and the final infection rate first falls then rises as the disclosing threshold increases, with the inflection point being at 0.8 (Figure 4a); the amount of final information per capita and the duration of government intervention both increase monotonically with the disclosing threshold (Figure 4b,e); the number of people infected after government intervention, the number of people infected by a panic after government intervention and the number of uninfected people remaining at the time of government intervention all negatively correlate with the disclosing threshold (Figure 4c,d,f).\nOne of the fundamental reasons for the government to balance speed and accuracy is the precipitous fall in the marginal contribution of accuracy as the disclosing threshold exceeds a certain “point”, while speed hardly affects the final infection rate. The following is a detailed analysis on the effects of both accuracy and speed.\nWith respect to accuracy, the effect comes from two perspectives: (1) accurate information lowers the infection from panic in a healthy panic-prone population (Figure 4d); (2) accurate information changes the behavior routine in the panic-prone population, which reduces the spread of the disease. Both (1) and (2) are in play until the disclosing threshold exceeds 0.5 (XI in our example); after that, there is no longer a panic-prone group, neither is infection from panic, which explains the precipitous fall in the marginal contribution of accuracy.\nWhen it comes to speed, the effect comes from two perspectives as well: (1) the time the government spends on censoring and screening information, which we call a lag; (2) the number of remaining uninfected people at the time of government disclosing information. The more accurate information the government seeks, the longer the lag (Figure 4e) and the fewer uninfected people remain at the time of disclosing (Figure 4f). Notice that both have roughly the same slope with respect to disclosing threshold, which explains a roughly constant marginal cost of pursuing accuracy.\nThe combination of constant marginal costs and abrupt fall in marginal benefits leads to an inflection point in disclosing threshold, which explains the heterogeneity in the distribution of optimal disclosing threshold: as disclosing threshold exceeds individual threshold, a sudden fall of benefits occurs, which theoretically makes the optimal disclosing threshold slightly greater than the individual threshold, which explains what we discussed above that the intervals in which the optimal disclosing threshold mostly lays differ.\nIn most cases, the government’s premature disclosing of inaccurate information will contaminate the overall network, while obsession with accuracy may have the government miss the disclosure window before too many people are infected, which is intolerable to government who requires a low infection rate. Thus, there is a trade-off."}

PMC:7795931 / 26835-32184 JSONTXT

Annnotations TAB JSON ListView MergeView

LitCovid-PubTator

LitCovid-PD-HP

LitCovid-sentences

PMC:7795931 / 26835-32184 JSON TXT