Updates from Overleaf

main.tex

@@ -97,7 +97,7 @@
  and
  \end{center}\\
 
- \textit{What would life be if we had no courage to attempt anything?} - Vincent Van Gogh\\
+ \textit{What would life be if we had (fear and) no courage to attempt anything?} - Vincent Van Gogh\\
 
  \end{tabularx}
 
@@ -219,18 +219,18 @@ \chapter{Introduction: The Two Cultures of Social Science}
 
 % XXX: to be improved
 In the more traditional social science and computational social science cultures, research is carried out to discover causal effects and model them. 
-For instance, propaganda and mass communication studies \cite{XXX} tries to understand the culture, time, authors, recipients in a non-invasive manner using the messages exchanged, and persuasion studies where the persuasion strategy present in the content is identified and correlated with (un)successful efforts of persuasion \cite{XXX}. 
+For instance, propaganda and mass communication studies \cite{mcquail1987mass,krippendorff2018content,lasswell1948structure,lasswell1971propaganda} try to understand the culture, time, authors, recipients in a non-invasive manner using the messages exchanged, and persuasion studies \cite{petty1981effects,chaiken1980heuristic} where the persuasion strategy present in the content is identified and correlated with (un)successful efforts of persuasion. 
 
 
 
 A common theme that runs through both research cultures in behavioral sciences is the intent to control behavior. Explanation and prediction are intermediate steps to control and hence optimize behavior. Optimizing behavior means to fulfill the communicator's objectives by controlling the other six parts of the communication process (Fig.~\ref{fig:factors-of-communication}). Due to the problem space being large, the solution needs a general understanding of human behavior as opposed to being domain-specific. 
 
 
-The characteristic that marks the digital age is the prevalence of human behavioral data in huge repositories. This data is big (allowing to model heterogeneity), always-on (allowing to look in the past as well as live measurements), observational (as opposed to reactive), but also incomplete (does not capture all that is happening everywhere everytime in a single repository) and algorithmically confounded (generated as a byproduct of an engineering process with a goal) \cite{salganik2019bit}. While the predictive culture has tried to make use of some of this data in the form of social media datasets like Twitter \cite{tumasjan2010predicting,asur2010predicting} and Instagram \cite{XXX}, Google trends \cite{choi2012predicting,carriere2013nowcasting}, Wikipedia \cite{generous2014global,de2021general,mestyan2013early}, shopping websites \cite{krumme2013predictability,de2015unique} and other data sources \cite{brockmann2006scaling,song2010limits,miritello2013limited}, these efforts are limited, in the sense of being dependent on one or a few chosen platforms, able to answer a limited set of questions, and restricted by access to private data. We want a model that can understand (predict and explain) human behavior as opposed to modeling a particular effect (retweet prediction) on a particular platform (\textit{e.g.} Twitter) for a certain type of users.
-This problem carries parallels with the problem being solved in the natural language processing (NLP) community, where supervised models in NLP are limited by the amount of supervision available and being able to answer one question (for which the supervised model was trained). The problem was solved by developing Large Language Models (LLMs), which are general purpose models capable of \textit{understanding language}, and hence can solve natural language tasks like sentiment analysis, question answering, email generation, and language translation in zero-shot (\textit{i.e.} without needing any explicit training for that task) \cite{XXX}.
+The characteristic that marks the digital age is the prevalence of human behavioral data in huge repositories. This data is big (allowing to model heterogeneity), always-on (allowing to look in the past as well as live measurements), observational (as opposed to reactive), but also incomplete (does not capture all that is happening everywhere everytime in a single repository) and algorithmically confounded (generated as a byproduct of an engineering process with a goal) \cite{salganik2019bit}. While the predictive culture has tried to make use of some of this data in the form of social media datasets like Twitter \cite{tumasjan2010predicting,asur2010predicting} and Instagram \cite{kim2020multimodal}, Google trends \cite{choi2012predicting,carriere2013nowcasting}, Wikipedia \cite{generous2014global,de2021general,mestyan2013early}, shopping websites \cite{krumme2013predictability,de2015unique} and other data sources \cite{brockmann2006scaling,song2010limits,miritello2013limited}, these efforts are limited, in the sense of being dependent on one or a few chosen platforms, able to answer a limited set of questions, and restricted by access to private data. We want a model that can understand (predict and explain) human behavior as opposed to modeling a particular effect (retweet prediction) on a particular platform (\textit{e.g.} Twitter) for a certain type of users.
+This problem carries parallels with the problem being solved in the natural language processing (NLP) community, where supervised models in NLP are limited by the amount of supervision available and being able to answer one question (for which the supervised model was trained). The problem was solved by developing Large Language Models (LLMs), which are general purpose models capable of \textit{understanding language}, and hence can solve natural language tasks like sentiment analysis, question answering, email generation, and language translation in zero-shot (\textit{i.e.} without needing any explicit training for that task) \cite{devlin2018bert,brown2020language,radford2018improving,raffel2020exploring,radford2019language}.
 
 
-Similarly, how do we develop a model capable of understanding behavior in general? With the intent to answer this question, we take motivation from LLMs where the idea is to train a model on a data-rich task. The task chosen to train LLMs is the next-word prediction, and the dataset is the text collected from the entire internet. We leverage the human behavior repositories available on the internet for this general-purpose human behavior model. The format of this data is the general communication model shown in Fig.~\ref{fig:factors-of-communication} consisting of communicator, message, time of message, channel, receiver, time of receipt, and effect. Due to the incomplete nature of behavioral repositories, all the factors are usually not always available. However, a subset is always available, and we show that the data scale helps make a general behavior understanding model \cite{khandelwal2023large}. We show that the model is capable of predicting behavior, explaining it, and also generating message to bring about certain behavior \cite{khurana2023behavior,si2023long,khandelwal2023large}.
+Similarly, how do we develop a model capable of understanding behavior \textit{in general}? With the intent to answer this question, we take motivation from LLMs where the idea is to train a model on a data-rich task. The task chosen to train LLMs is the next-word prediction, and the dataset is the text collected from the entire internet. We leverage the human behavior repositories available on the internet for this general-purpose human behavior model. The format of this data is the general communication model shown in Fig.~\ref{fig:factors-of-communication} consisting of communicator, message, time of message, channel, receiver, time of receipt, and effect. Due to the incomplete nature of behavioral repositories, all the factors are usually not always available. However, a subset is always available, and we show that the data scale helps make a general behavior understanding model \cite{khandelwal2023large}. We show that the model is capable of predicting behavior, explaining it, and also generating message to bring about certain behavior \cite{khurana2023behavior,si2023long,khandelwal2023large}.
 
 
 
@@ -1232,7 +1232,7 @@ \subsubsection{A few examples of the stories generated using our method}
 
 
 
- \noindent3.~\url{https://www.youtube.com/watch?v=PJlHiQJBDMw}: The advertisement for the company Sherwin-Williams opens on a kitchen table strewn with shells and wicker baskets, with two glasses of iced water beside them. A vase with a blue pattern sits in the foreground, and a person holds up a phone with the Sherwin-Williams logo on the screen. A girl appears from behind a white sheet, peeking out of a white tent as if to signify the timelessness and neutrality of this color. The voice-over begins, as the camera pans to a living room with a staircase, and then to a dining room with a white table, chairs, and a white vase. The words "Color of the Month: Shell White, Sherwin-Williams" appear on the screen, as the camera zooms in on the vase. The words are followed by \"Our app makes it a snap,\" referring to Color Snap, the company's new way of painting a home. The advertisement ends with the Sherwin-Williams logo, emphasizing the company's commitment to excellence in home painting. This advertisement was promoting the company's color of the month, Shell White. \\
+ \noindent3.~\url{https://www.youtube.com/watch?v=PJlHiQJBDMw}: The advertisement for the company Sherwin-Williams opens on a kitchen table strewn with shells and wicker baskets, with two glasses of iced water beside them. A vase with a blue pattern sits in the foreground, and a person holds up a phone with the Sherwin-Williams logo on the screen. A girl appears from behind a white sheet, peeking out of a white tent as if to signify the timelessness and neutrality of this color. The voice-over begins, as the camera pans to a living room with a staircase, and then to a dining room with a white table, chairs, and a white vase. The words ``Color of the Month: Shell White, Sherwin-Williams" appear on the screen, as the camera zooms in on the vase. The words are followed by \"Our app makes it a snap,\" referring to Color Snap, the company's new way of painting a home. The advertisement ends with the Sherwin-Williams logo, emphasizing the company's commitment to excellence in home painting. This advertisement was promoting the company's color of the month, Shell White. \\
 
  \noindent4.~\url{https://www.youtube.com/watch?v=CDjBIt70fp4}: The story began with a green light glowing in the dark, symbolizing the presence of a powerful technology that can change the way we work. This technology was an advanced graphics card, the NVIDIA Quadro FX 1700. It was compared side-by-side with its successor, the Quadro 2000, and it was clear that the Quadro 2000 was far more powerful. The Quadro FX 1700 had a qt - x700 vs qt - x700 capacity, while the Quadro 2000 had a green light that shone brighter and further than before. As the comparison was being made, the results were clear: the Quadro 2000 was the superior product. This advertisement for the company NVIDIA showcased the power of the Quadro 2000, and the improved performance it could bring to an organization. The advertisement concluded with a green light, signaling that NVIDIA had the answer to improving workflows. The product the advertisement was about was the NVIDIA Quadro 2000." \\
 

ref.bib

@@ -2401,6 +2401,40 @@ @techreport{r:86
  type = {Technical Report},
  institution = {Dept.\ of Computer Science, Stanford Univ.}
 }
+@inproceedings{kim2020multimodal,
+ title={Multimodal post attentive profiling for influencer marketing},
+ author={Kim, Seungbae and Jiang, Jyun-Yu and Nakada, Masaki and Han, Jinyoung and Wang, Wei},
+ booktitle={Proceedings of The Web Conference 2020},
+ pages={2878--2884},
+ year={2020}
+}
+@book{krippendorff2018content,
+ title={Content analysis: An introduction to its methodology},
+ author={Krippendorff, Klaus},
+ year={2018},
+ publisher={Sage publications}
+}
+@book{mcquail1987mass,
+ title={Mass communication theory: An introduction},
+ author={McQuail, Denis},
+ year={1987},
+ publisher={Sage Publications, Inc}
+}
+@article{radford2019language,
+ title={Language models are unsupervised multitask learners},
+ author={Radford, Alec and Wu, Jeffrey and Child, Rewon and Luan, David and Amodei, Dario and Sutskever, Ilya and others},
+ journal={OpenAI blog},
+ volume={1},
+ number={8},
+ pages={9},
+ year={2019}
+}
+@article{radford2018improving,
+ title={Improving language understanding by generative pre-training},
+ author={Radford, Alec and Narasimhan, Karthik and Salimans, Tim and Sutskever, Ilya and others},
+ year={2018},
+ publisher={OpenAI}
+}
 @inproceedings{radford2023robust,
  title = {Robust speech recognition via large-scale weak supervision},
  author = {Radford, Alec and Kim, Jong Wook and Xu, Tao and Brockman, Greg and McLeavey, Christine and Sutskever, Ilya},
@@ -3249,7 +3283,7 @@ @inproceedings{diba2020large
  pages = {593--610},
  organization = {Springer}
 }
-@software{breakthrough-pyscenedetect,
+@misc{breakthrough-pyscenedetect,
  title = {PySceneDetect: Video Scene Cut Detection and Analysis Tool},
  author = {Breakthrough},
  year = 2023,