4 ms·

Maps of subjective feelings

40 points by cheeaun 8y ago

johnalamos 8y ago
?????
johnalamosisi 8y ago
Is this a joke
scienterrific 8y ago
This is the most bullshit I've seen packed into a scientific publication wrapper in quite a while. Look at this graph: http://www.pnas.org/content/pnas/early/2018/08/27/1807390115/F2.large.jpg http://www.pnas.org/content/pnas/early/2018/08/27/1807390115... What? And this image here is like an uncanny valley of robot perception trying to comprehend the vagaries of sensation: http://www.pnas.org/content/pnas/early/2018/08/27/1807390115/F3.large.jpg http://www.pnas.org/content/pnas/early/2018/08/27/1807390115... Why thank you, robo-droid! Very intuitive! I'm putting this on the refrigerator! I feel like I just accidentally read an algorithmically generated ad-tech article from a chum bucket click-funnel content farm.
- snapspans 8y ago
  Still working out how to read it (from just a glance at your link it looks like it ought to be investigated, or interpreted): Caption: "Fig. 2. The feeling space. (Upper) Two-dimensional map of the feeling space based on the sorting task average distance matrix between items arranged by t-SNE and clustered with DBSCAN. Color coding indicates cluster structure; gray feelings do not belong clearly to any cluster. Colored items with black edge are DBSCAN border elements. To retain the information of the distance matrix the closest three items for each node are connected with lines. Thick dark lines are showing distances that belong to the top 33rd percentile of the visualized lines (i.e., the closest items). (Lower) Heatmaps showing how strongly each basic dimension of subjective experience is associated with each discrete feeling at each location of the feeling space. Color coding shows the relative intensity as median z-score (as in Fig. 1) from high (red) to low (blue)." Edit: The second image you linked would be perfect for the refrigerator. Great observation. (I'm not sure if I'm feeling "sadness" or "laughing".)
- ivraatiems 8y ago
  Can you tell us a little bit about your scientific background which would help us understand your qualifications to make that statement? I'm really tired of non-subject-matter experts on HN coming in and declaring things that require very specific expertise bullshit because they didn't immediately understand or agree with them on a first pass. I'm not saying you're doing that, but that's an awful lot like what it seems like given how argumentative your comment is.
  - scienterrific 8y ago
    I can tell you how unscientific the process of this study was: A total of 1,026 participants took part in online surveys where we assessed: (i) for each feeling, the intensity of four hypothesized basic dimensions (ii) subjectively experienced similarity of the 100 feelings, and (iii) topography of bodily sensations associated with each feeling. They mechanical turked some online surveys, and asked non-experts for this data, on the basis that it’s subjective data. Then they performed some canned statistical analysis du jour on it, so that the paper would have some catchy data visualizations, and filled for 1,000 words. So, first and foremost, this is a free-association study, with all the rigor of a Freudian interpretation for responses to a Rorschach test. But beyond that, why would anyone hold the opinion that this is science? Online surveys for 1,000 people, to gather opinions on the best way to associate words in English? That’s what this research accomplished. And lastly, let’s imagine the potential applications for these results. How will they guide us to better predictions in the future? If I look at the information collected here, can I draw more accurate conclusions about the world, and improve practical decisions, now that I’ve been informed by these results? No. No I can’t. This is the kind of analysis that simply trusts the respondant to complete an online survey in good faith. You can’t run the world like that. You can’t drop ten cents on a 30 minute multiple choice questionaire, and use that as your foundation for new insights. They basically asked people to generate a tackling dummy data set for them by completing a survey, so that they could run a sort algorithm on it, and apply a graph to an Excel spreadsheet of answer counts. Except they did it in an academic setting, so that some how validates these results with defacto authority? Not in my world.
    - 8y ago

40 points by cheeaun 8y ago

johnalamos 8y ago

?????

johnalamosisi 8y ago

Is this a joke

scienterrific 8y ago

This is the most bullshit I've seen packed into a scientific publication wrapper in quite a while. Look at this graph: http://www.pnas.org/content/pnas/early/2018/08/27/1807390115/F2.large.jpg http://www.pnas.org/content/pnas/early/2018/08/27/1807390115... What? And this image here is like an uncanny valley of robot perception trying to comprehend the vagaries of sensation: http://www.pnas.org/content/pnas/early/2018/08/27/1807390115/F3.large.jpg http://www.pnas.org/content/pnas/early/2018/08/27/1807390115... Why thank you, robo-droid! Very intuitive! I'm putting this on the refrigerator! I feel like I just accidentally read an algorithmically generated ad-tech article from a chum bucket click-funnel content farm.

snapspans 8y ago
Still working out how to read it (from just a glance at your link it looks like it ought to be investigated, or interpreted): Caption: "Fig. 2. The feeling space. (Upper) Two-dimensional map of the feeling space based on the sorting task average distance matrix between items arranged by t-SNE and clustered with DBSCAN. Color coding indicates cluster structure; gray feelings do not belong clearly to any cluster. Colored items with black edge are DBSCAN border elements. To retain the information of the distance matrix the closest three items for each node are connected with lines. Thick dark lines are showing distances that belong to the top 33rd percentile of the visualized lines (i.e., the closest items). (Lower) Heatmaps showing how strongly each basic dimension of subjective experience is associated with each discrete feeling at each location of the feeling space. Color coding shows the relative intensity as median z-score (as in Fig. 1) from high (red) to low (blue)." Edit: The second image you linked would be perfect for the refrigerator. Great observation. (I'm not sure if I'm feeling "sadness" or "laughing".)
ivraatiems 8y ago
Can you tell us a little bit about your scientific background which would help us understand your qualifications to make that statement? I'm really tired of non-subject-matter experts on HN coming in and declaring things that require very specific expertise bullshit because they didn't immediately understand or agree with them on a first pass. I'm not saying you're doing that, but that's an awful lot like what it seems like given how argumentative your comment is.
- scienterrific 8y ago
  I can tell you how unscientific the process of this study was: A total of 1,026 participants took part in online surveys where we assessed: (i) for each feeling, the intensity of four hypothesized basic dimensions (ii) subjectively experienced similarity of the 100 feelings, and (iii) topography of bodily sensations associated with each feeling. They mechanical turked some online surveys, and asked non-experts for this data, on the basis that it’s subjective data. Then they performed some canned statistical analysis du jour on it, so that the paper would have some catchy data visualizations, and filled for 1,000 words. So, first and foremost, this is a free-association study, with all the rigor of a Freudian interpretation for responses to a Rorschach test. But beyond that, why would anyone hold the opinion that this is science? Online surveys for 1,000 people, to gather opinions on the best way to associate words in English? That’s what this research accomplished. And lastly, let’s imagine the potential applications for these results. How will they guide us to better predictions in the future? If I look at the information collected here, can I draw more accurate conclusions about the world, and improve practical decisions, now that I’ve been informed by these results? No. No I can’t. This is the kind of analysis that simply trusts the respondant to complete an online survey in good faith. You can’t run the world like that. You can’t drop ten cents on a 30 minute multiple choice questionaire, and use that as your foundation for new insights. They basically asked people to generate a tackling dummy data set for them by completing a survey, so that they could run a sort algorithm on it, and apply a graph to an Excel spreadsheet of answer counts. Except they did it in an academic setting, so that some how validates these results with defacto authority? Not in my world.
  - 8y ago