skip to main content

Title: CASM: A Deep-Learning Approach for Identifying Collective Action Events with Text and Image Data from Social Media
Protest event analysis is an important method for the study of collective action and social movements and typically draws on traditional media reports as the data source. We introduce collective action from social media (CASM)—a system that uses convolutional neural networks on image data and recurrent neural networks with long short-term memory on text data in a two-stage classifier to identify social media posts about offline collective action. We implement CASM on Chinese social media data and identify more than 100,000 collective action events from 2010 to 2017 (CASM-China). We evaluate the performance of CASM through cross-validation, out-of-sample validation, and comparisons with other protest data sets. We assess the effect of online censorship and find it does not substantially limit our identification of events. Compared to other protest data sets, CASM-China identifies relatively more rural, land-related protests and relatively few collective action events related to ethnic and religious conflict.
Authors:
;
Award ID(s):
1831481
Publication Date:
NSF-PAR ID:
10182271
Journal Name:
Sociological Methodology
Volume:
49
Issue:
1
Page Range or eLocation-ID:
1 to 57
ISSN:
0081-1750
Sponsoring Org:
National Science Foundation
More Like this
  1. Web-based interactions allow agents to coordinate and to take actions (change state) jointly, i.e., to participate in collective action such as a protest, facilitating spread of contagion to large groups within networked populations. In game theoretic contexts, coordination requires that agents share common knowledge about each other. Common knowledge emerges within a group when each member knows the states and the types (preferences) of the other members, and critically, each member knows that everyone else has this information. Hence, these models of common knowledge and coordination on communication networks are fundamentally different from influence-based unilateral contagion models, such as those devised by Granovetter and Centola. Common knowledge arises in many settings in practice, yet there are few operational models that can be used to compute contagion dynamics. Moreover, these models utilize different mechanisms for driving contagion. We evaluate the three mechanisms of a common knowledge model that can represent web-based communication among groups of people on Facebook. We evaluate these mechanisms on five social (media) networks with wide-ranging properties. We demonstrate that different mechanisms can produce widely varying behaviors in terms of the extent of contagion spreading and the speed of contagion transmission.
  2. Protest is a collective action problem and can be modeled as a coordination game in which people take an action with the potential to achieve shared mutual benefits. In game-theoretic contexts, successful coordination requires that people know each others' willingness to participate, and that this information is common knowledge among a sufficient number of people. We develop an agent-based model of collective action that was the first to combine social structure and individual incentives. Another novel aspect of the model is that a social network increases in density (i.e., new graph edges are formed) over time. The model studies the formation of common knowledge through local interactions and the characterizing social network structures. We use four real-world, data-mined social networks (Facebook, Wikipedia, email, and peer-to-peer networks) and one scale-free network, and conduct computational experiments to study contagion dynamics under different conditions.
  3. Bae, K-H ; Feng, B ; Kim, S ; Lazarova-Molnar, S ; Zheng, Z ; Roeder, T ; Thiesing, R. (Ed.)
    Protest is a collective action problem and can be modeled as a coordination game in which people take an action with the potential to achieve shared mutual benefits. In game-theoretic contexts, successful coordination requires that people know each others’ willingness to participate, and that this information is common knowledge among a sufficient number of people. We develop an agent-based model of collective action that was the first to combine social structure and individual incentives. Another novel aspect of the model is that a social network increases in density (i.e., new graph edges are formed) over time. The model studies the formation of common knowledge through local interactions and the characterizing social network structures. We use four real-world, data-mined social networks (Facebook, Wikipedia, email, and peer-to-peer networks) and one scale-free network, and conduct computational experiments to study contagion dynamics under different conditions.
  4. Web-based interactions allow agents to coordinate and to take actions (change state) jointly, i.e., to participate in collective ac- tion such as a protest, facilitating spread of contagion to large groups within networked populations. In game theoretic contexts, coordination requires that agents share common knowledge about each other. Common knowledge emerges within a group when each member knows the states and the types (preferences) of the other members, and critically, each member knows that everyone else has this information. Hence, these models of common knowledge and coordination on communication networks are fundamentally different from influence-based unilateral contagion models, such as those devised by Granovetter and Centola. Common knowledge arises in many settings in practice, yet there are few operational models that can be used to compute contagion dynamics. Moreover, these models utilize different mechanisms for driving contagion. We evaluate the three mechanisms of a common knowledge model that can represent web-based communication among groups of people on Facebook. We evaluate these mechanisms on ve social (media) networks with wide-ranging properties.We demonstrate that di erent mechanisms can produce widely varying behaviors in terms of the extent of contagion spreading and the speed of contagion transmission.
  5. Obeid, I. ; Selesnik, I. ; Picone, J. (Ed.)
    The Neuronix high-performance computing cluster allows us to conduct extensive machine learning experiments on big data [1]. This heterogeneous cluster uses innovative scheduling technology, Slurm [2], that manages a network of CPUs and graphics processing units (GPUs). The GPU farm consists of a variety of processors ranging from low-end consumer grade devices such as the Nvidia GTX 970 to higher-end devices such as the GeForce RTX 2080. These GPUs are essential to our research since they allow extremely compute-intensive deep learning tasks to be executed on massive data resources such as the TUH EEG Corpus [2]. We use TensorFlow [3] as the core machine learning library for our deep learning systems, and routinely employ multiple GPUs to accelerate the training process. Reproducible results are essential to machine learning research. Reproducibility in this context means the ability to replicate an existing experiment – performance metrics such as error rates should be identical and floating-point calculations should match closely. Three examples of ways we typically expect an experiment to be replicable are: (1) The same job run on the same processor should produce the same results each time it is run. (2) A job run on a CPU and GPU should producemore »identical results. (3) A job should produce comparable results if the data is presented in a different order. System optimization requires an ability to directly compare error rates for algorithms evaluated under comparable operating conditions. However, it is a difficult task to exactly reproduce the results for large, complex deep learning systems that often require more than a trillion calculations per experiment [5]. This is a fairly well-known issue and one we will explore in this poster. Researchers must be able to replicate results on a specific data set to establish the integrity of an implementation. They can then use that implementation as a baseline for comparison purposes. A lack of reproducibility makes it very difficult to debug algorithms and validate changes to the system. Equally important, since many results in deep learning research are dependent on the order in which the system is exposed to the data, the specific processors used, and even the order in which those processors are accessed, it becomes a challenging problem to compare two algorithms since each system must be individually optimized for a specific data set or processor. This is extremely time-consuming for algorithm research in which a single run often taxes a computing environment to its limits. Well-known techniques such as cross-validation [5,6] can be used to mitigate these effects, but this is also computationally expensive. These issues are further compounded by the fact that most deep learning algorithms are susceptible to the way computational noise propagates through the system. GPUs are particularly notorious for this because, in a clustered environment, it becomes more difficult to control which processors are used at various points in time. Another equally frustrating issue is that upgrades to the deep learning package, such as the transition from TensorFlow v1.9 to v1.13, can also result in large fluctuations in error rates when re-running the same experiment. Since TensorFlow is constantly updating functions to support GPU use, maintaining an historical archive of experimental results that can be used to calibrate algorithm research is quite a challenge. This makes it very difficult to optimize the system or select the best configurations. The overall impact of all of these issues described above is significant as error rates can fluctuate by as much as 25% due to these types of computational issues. Cross-validation is one technique used to mitigate this, but that is expensive since you need to do multiple runs over the data, which further taxes a computing infrastructure already running at max capacity. GPUs are preferred when training a large network since these systems train at least two orders of magnitude faster than CPUs [7]. Large-scale experiments are simply not feasible without using GPUs. However, there is a tradeoff to gain this performance. Since all our GPUs use the NVIDIA CUDA® Deep Neural Network library (cuDNN) [8], a GPU-accelerated library of primitives for deep neural networks, it adds an element of randomness into the experiment. When a GPU is used to train a network in TensorFlow, it automatically searches for a cuDNN implementation. NVIDIA’s cuDNN implementation provides algorithms that increase the performance and help the model train quicker, but they are non-deterministic algorithms [9,10]. Since our networks have many complex layers, there is no easy way to avoid this randomness. Instead of comparing each epoch, we compare the average performance of the experiment because it gives us a hint of how our model is performing per experiment, and if the changes we make are efficient. In this poster, we will discuss a variety of issues related to reproducibility and introduce ways we mitigate these effects. For example, TensorFlow uses a random number generator (RNG) which is not seeded by default. TensorFlow determines the initialization point and how certain functions execute using the RNG. The solution for this is seeding all the necessary components before training the model. This forces TensorFlow to use the same initialization point and sets how certain layers work (e.g., dropout layers). However, seeding all the RNGs will not guarantee a controlled experiment. Other variables can affect the outcome of the experiment such as training using GPUs, allowing multi-threading on CPUs, using certain layers, etc. To mitigate our problems with reproducibility, we first make sure that the data is processed in the same order during training. Therefore, we save the data from the last experiment and to make sure the newer experiment follows the same order. If we allow the data to be shuffled, it can affect the performance due to how the model was exposed to the data. We also specify the float data type to be 32-bit since Python defaults to 64-bit. We try to avoid using 64-bit precision because the numbers produced by a GPU can vary significantly depending on the GPU architecture [11-13]. Controlling precision somewhat reduces differences due to computational noise even though technically it increases the amount of computational noise. We are currently developing more advanced techniques for preserving the efficiency of our training process while also maintaining the ability to reproduce models. In our poster presentation we will demonstrate these issues using some novel visualization tools, present several examples of the extent to which these issues influence research results on electroencephalography (EEG) and digital pathology experiments and introduce new ways to manage such computational issues.« less