Determining Subjective Bias in Text through Linguistically Informed Transformer based Multi-Task Network

2021 ◽  
Author(s):  
Manjira Sinha ◽  
Tirthankar Dasgupta
Keyword(s):  
1947 ◽  
Vol 93 (390) ◽  
pp. 31-42 ◽  
Author(s):  
E. L. Hutton

In this paper are given the histories of ten patients who have been treated by means of leucotomy, particular attention being paid to the study of character and personality traits.The records are compiled from the information given by a relative or relatives, supplemented in some cases by data supplied by the patient as well. This method undoubtedly has serious disadvantages, for the information must inevitably be dependent upon the informant's veracity, powers of observation and description, memory, and conscious and unconscious subjective bias. In every case the interview took place at the patient's home, so that the interviewer could obtain a personal impression of the patient's home environment. Furthermore, the information regarding the pre-morbid personality was only obtained after the illness, operation, and recovery, and in some cases it was difficult to ascertain when the morbid changes actually began.


2019 ◽  
Author(s):  
Billy J. Andrews ◽  
Jennifer J. Roberts ◽  
Zoe K. Shipton ◽  
Sabina Bigi ◽  
Maria C. Tartarello ◽  
...  

Abstract. The characterisation of natural fracture networks using outcrop analogues is important in understanding sub-surface fluid flow and rock mass characteristics in fractured lithologies. It is well known from decision-sciences that subjective bias significantly impacts the way data is gathered and interpreted. This study investigates the impact of subjective bias on fracture data collected using four commonly used approaches (linear scanlines, circular scanlines, topology sampling and window sampling) both in the field and in workshops using field photographs. Considerable variability is observed between each participant's interpretation of the same scanline, and this variability is seen regardless of geological experience. Geologists appear to be either focussing on the detail or focussing on gathering larger volumes of data, and this innate personality trait affects the recorded fracture network attributes. As a result, fracture statistics derived from the field data and which are often used as inputs for geological models, can vary considerably between different geologists collecting data from the same scanline. Additionally, the personal bias of geologists collecting the data affects the size (minimum length of linear scanlines, radius of circular scanlines or area of a window sample) required of the scanline that is needed to collect a statistically representative amount of data. We suggest protocols to recognise, understand and limit the effect of subjective bias on fracture data biases during data collection.


Pain Medicine ◽  
2019 ◽  
Vol 20 (7) ◽  
pp. 1347-1354
Author(s):  
Hayato Shigetoh ◽  
Michihiro Osumi ◽  
Shu Morioka

Abstract Background Manual traction is used for pain relief, but it is not clear whether the pain relief effect of manual traction is due to sensitivity or to subjective bias. The differences between manual traction and touch have also been unclear. Objectives We used signal detection theory to investigate whether manual traction and touch were effective for pain relief, and we compared the pain relief effect between manual traction and touch. Design Repeated measures and single blinding. Methods Twenty healthy adult volunteers performed an intensity judgment task immediately before and after each intervention. The intervention was either manual traction or touch for 10 minutes. We measured the intensity judgment task’s signal detection measures of hit rates, false alarm rates, sensitivity (d'), and response bias (C) in an Aδ fiber–mediated pain condition and C fiber–mediated pain condition. Results Manual traction did not provide a significant level of change, but its effect sizes differed. In our comparison of the effect sizes, manual traction tended to reduce the hit rate and altered the sensitivity value rather than the response bias in Aδ fiber–mediated pain. There was no significant difference in the amount of change in the hit rate between touch and manual traction regarding Aδ fiber–mediated pain and C fiber–mediated pain. Conclusions In terms of effect sizes, manual traction was effective for the pain relief of the first pain by producing a change in pain sensitivity rather than by subjective bias. Manual traction reduced the first pain, whereas touch reduced the first pain and second pain.


2012 ◽  
Vol 5 (1) ◽  
pp. 42-53
Author(s):  
Mark Titmarsh

The history of the painted image has involved various metamorphoses from cave, to architecture, to easel and most recently the radical hybridisation of expanded multimedia forms. Through each age the shedding of one aspect of the image; ritual, spirituality, portability, has resulted in a shift from sacred static images towards profane ephemeral events. This transformation has intensified over the last century where the repeated announcement of the death of painting has seen painting reborn as a mode of radical self-questioning. This paper takes an overview of painting's morphology by way of Martin Heidegger's discussion of picturing and representation in his essay “The Age of the World Picture” (1938). By focusing on the history of picturing, an understanding of “original aesthetics” and its apprehensions can be developed, for comparison with contemporary art in its “post-medium condition”. The new ontological paradigm of contemporary art therefore demands another discourse, a “post-aesthetics” that overcomes the subjective bias of modern philosophical aesthetics in favour of a primary relationship to things and their mode of presence in the world. As a result, contemporary expanded painting is shown to be a radical revision of art, a moment of ontological “presencing” favouring spatial environments and temporal events that reveal “what is” and “what matters” in a contemporary techno-scientific age. As a result ontological aesthetics could indicate what is good at the boundary between people, communities, technology and the earth, in short, a politics of being. In the politics of being, which is most appropriate to the expanded work of art, politics is linked to the polis, the place where culture emerges into a world, and where things unfold according to the sense of possibility that a world grants. Art is intimately involved in an alternative politics of being when it poses another way of dealing with beings, people, objects, and things, that result in new economies and temporalities indicating the possibility of something beyond our habitual understanding of the world. Expanded painting as the exemplar of contemporary art is an ontological cut in our understanding of art and the world. It slices through the contemporary understanding of presence and delivers a monstrous thickness, no longer supported by a surface as substrate, but instead compelled by the phenomenal experiences of the world, and is thereby multiplied exponentially and existentially. Santrauka Tapyto atvaizdo istorija apima įvairias metamorfozes, pradedant urvais ir baigiant architektūra, molbertais, o pastaruoju metu – radikalia išplėstinių multimedijų formų hibridizacija. Bet kuriame amžiuje nunykdavo tam tikras atvaizdo aspektas; ritualai, dvasingumas, portatyvumas sakralius statiškus atvaizdus pakeitė profaniškais trumpalaikiais įvykiais. Ši transformacija suintensyvėjo pastarajame amžiuje, kai buvo pakartotinai paskelbta, esą tapyba mirė, atgimstant tapybai kaip radikaliai saviklausai. Šiame straipsnyje pateikiama tapybos morfologijos apžvalga, remiantis Martino Heideggerio apmąstymais. Susitelkiant į atvaizdų istoriją, gali būti išplėtota „originaliosios estetikos“ ir jos suvokimo samprata, lyginant su šiuolaikiniu menu ir jo „postmedijine būkle“. Todėl nauja šiuolaikinio meno ontologinė paradigma reikalauja kito – „postestetikos“ – diskurso, įveikiančio subjektyvų moderniosios filosofinės estetikos šališkumą pirminio santykio su daiktais ir su jų buvimo pasaulyje būdo atžvilgiu. Iš to išplaukia, kad šiuolaikinė išplėstinė tapyba pristatoma kaip radikalus meno peržiūrėjimas, ontologinio „esamumo“ akimirka, palaikanti erdvines aplinkas ir laikinius įvykius, atveriančius „tai, kas yra“ ir „tai, kas vyksta“ šiuolaikiniame technikos ir mokslo amžiuje. Tad ontologinė estetika gali nurodyti, kas yra gera ties žmonių, bendruomenių, technologijų ir Žemės ribomis, trumpai tariant, atskleisti būties politiką. Šioji yra pati tinkamiausia išplėstiniams meno kūriniams; politika yra susieta su polis – ta vieta, kurioje kultūra iškyla į pasaulį ir kurioje pagal pasaulio teikiamą galimumą skleidžiasi daiktai. Menas yra glaudžiai įtraukiamas į alternatyviąją būties politiką, iškeldamas naują būtybių, žmonių, objektų ir daiktų santykiavimo būdą. Tai lemia naujose ekonomikose ir laikiškumuose nurodomą galimybę kažko, kas yra anapus mūsų pasaulio įprastinio supratimo. Išplėstinė tapyba kaip šiuolaikinio meno pavyzdys yra ontologinis įtrūkis mums suprantant meną ir pasaulį. Ji aptinkama suprantant šiuolaikinį buvimą, yra labai reikšminga ir nebėra palaikoma paviršiaus kaip substrato; veikiau įtikinamų fenomenalių pasaulio patirčių, taip dauginamų išoriškai ir egzistenciškai.


2018 ◽  
Vol 13 (02) ◽  
pp. 201-222
Author(s):  
Lub Lyna Nabilata

Cultural optics will always appear in every debate about feminism, as well as certain interpretive tendencies (read: pre-text) also involved (enveloped) and even come into play in them. Therefore, the emergence of different views even somewhat "biased" is considered normal. People in discussing feminism will not be able to position themselves really objectively without pretension, but can only maintain a distance from prejudices or “biases” that can unwittingly emerge. In Muslim feminist thought, they are still trapped in a crisis of interpretation and counter interpretation. This crisis arises because the methods or strategies used by opponents and supporters of gender equality in building and legitimizing each of their views are basically the same, namely by explaining certain parts of the text of the Qur’an or hadith that are appropriate and support their interests and views. The parts of the text are then considered as asl the most correct and original principles, which in turn tend to be interpreted unilaterally in accordance with their ideological interests and positions and at the same time eliminate unwanted meanings because they are contrary to their ideological interests. This eclectic reading model is caused by the inability to challenge the existing paradigm of reading the text, which does not consider the historical context, dialogical and communicative aspects of the text with its context, and its descriptive dimensions. In this article intending to criticize the interpretation of Fatima Mernissi, Mernissi seems to still have a subjective bias in assessing some of the problems of feminist interpretation regarding equality of men and women, and not occupying the core of the problem in the actual portion. As a result, this argument shows that Mernissi is still narrow in using her feminist approach, because the core feminist approach is actually sensitive to injustice and avoiding “bias” that can occur not only in gender issues, but can also occur in other areas related to sara (read: skin color, tribe, caste and others).


10.28945/3277 ◽  
2008 ◽  
Author(s):  
Peter Bednar ◽  
Christine Welch

What is normally described as bias? A possible definition comprises attempts to distort or mislead to achieve a certain perspective, i.e. subjective descriptions intended to mislead. If designers were able to exclude bias from informing systems, then this would maximize their effectiveness. This implicit conjecture appears to underpin much of the research in our field. However, in our efforts to support the evolution and design of informing systems, the way we think, communicate and conceptualize our efforts clearly influences our comprehension and consequently our agenda for design. Objectivity (an attempt to be neutral or transparent) is usually regarded as exclusion of bias. However, claims for objectivity do not, by definition, include efforts to inquire into and reflect over subjective values. Attempts to externalize the mindset of the subject do not arise as part of the description. When claims to objectivity are made, this rarely includes any effort to make subjective bias transparent. Instead, objectivity claims may be regarded as a denial of bias. We suggest that bias can be introduced into overt attempts to admit subjectivity. For example, where people are asked to give subjective opinion according to an artificially enforced scale of truth-falsity (bi-valued logic), they may find themselves coerced into statements of opinion that do not truly reflect the views they might have wished to express. People do not naturally respond to their environment with opinions limited to restricted scales; rather, they tend to use multi-valued, or para-consistent logic. This paper examines the impact of bias within attempts to establish communicative practice in human activity systems (informing systems).


1997 ◽  
Vol 36 (03) ◽  
pp. 191-200 ◽  
Author(s):  
G. T. H. Ellison ◽  
M. Holliday ◽  
E. Lucassen ◽  
H. E. Harris

The quality of maternal information contained within contemporary obstetric notes was investigated by abstracting data from the medical records of multiparous women who were admitted to a major city hospital in the South Thames Region. Potential sources of error were identified by comparing information recorded in different sections of each obstetric notes and within the obstetric notes of consecutive pregnancies. The format of the obstetric notes largely determined which variables were recorded and, to some extent, the accuracy of information collected. However, the quality of the data ultimately depended upon whether each variable was self-reported or directly measured. Self-reported variables were subject to selective omission and subjective bias, while measured variables were susceptible to inaccurate equipment and poor measurement practice. By interviewing a sample of midwives currently involved in antenatal care at the Trust it was possible to confirm that extensive variation in measurement and recording procedures routinely occurred.


1996 ◽  
Vol 82 (3) ◽  
pp. 731-734
Author(s):  
Charles D. Bailey

The existence of a pervasive and pronounced steepness bias in visually fitting a line to data in a scattergraph was investigated. 262 undergraduate business students were asked to fit lines visually to scattergraphs, to correspond to a least squares regression fit. These visually estimated lines strongly overestimated the steepness of the actual trend. As visual inspection of data is an important step for the detection of linear trends and outliers prior to regression analysis, this bias in subjective perception of the line may result in inappropriate deletion or retention of outliers, transferring the subjective bias into the results of least squares regression. In addition, textbooks on management accounting discuss the use of visually fitted lines to estimate cost behavior.


Sign in / Sign up

Export Citation Format

Share Document