Deeper neural networks are more difficult to train. We present a residual learning framework to ease the training of networks that are substantially deeper than those used previously. We explicitly reformulate the layers as learning residual functions with reference to the layer inputs, instead of learning unreferenced functions. We provide comprehensive empirical evidence showing that these residual networks are easier to optimize, and can gain accuracy from considerably increased depth. On the ImageNet dataset we evaluate residual nets with a depth of up to 152 layers—8× deeper than VGG nets [40] but still having lower complexity. An ensemble of these residual nets achieves 3.57% error on the ImageNet test set. This result won the 1st place on the ILSVRC 2015 classification task. We also present analysis on CIFAR-10 with 100 and 1000 layers. The depth of representations is of central importance for many visual recognition tasks. Solely due to our extremely deep representations, we obtain a 28% relative improvement on the COCO object detection dataset. Deep residual nets are foundations of our submissions to ILSVRC & COCO 2015 competitions1, where we also won the 1st places on the tasks of ImageNet detection, ImageNet localization, COCO detection, and COCO segmentation.

/pdf/deep-residual-learning-for-image-recognition-b75jg61qrq.pdf

Deep Residual Learning for Image Recognition

In this work we investigate the effect of the convolutional network depth on its accuracy in the large-scale image recognition setting. Our main contribution is a thorough evaluation of networks of increasing depth using an architecture with very small (3x3) convolution filters, which shows that a significant improvement on the prior-art configurations can be achieved by pushing the depth to 16-19 weight layers. These findings were the basis of our ImageNet Challenge 2014 submission, where our team secured the first and the second places in the localisation and classification tracks respectively. We also show that our representations generalise well to other datasets, where they achieve state-of-the-art results. We have made our two best-performing ConvNet models publicly available to facilitate further research on the use of deep visual representations in computer vision.

Very Deep Convolutional Networks for Large-Scale Image Recognition

Deeper neural networks are more difficult to train. We present a residual learning framework to ease the training of networks that are substantially deeper than those used previously. We explicitly reformulate the layers as learning residual functions with reference to the layer inputs, instead of learning unreferenced functions. We provide comprehensive empirical evidence showing that these residual networks are easier to optimize, and can gain accuracy from considerably increased depth. On the ImageNet dataset we evaluate residual nets with a depth of up to 152 layers---8x deeper than VGG nets but still having lower complexity. An ensemble of these residual nets achieves 3.57% error on the ImageNet test set. This result won the 1st place on the ILSVRC 2015 classification task. We also present analysis on CIFAR-10 with 100 and 1000 layers. 
The depth of representations is of central importance for many visual recognition tasks. Solely due to our extremely deep representations, we obtain a 28% relative improvement on the COCO object detection dataset. Deep residual nets are foundations of our submissions to ILSVRC & COCO 2015 competitions, where we also won the 1st places on the tasks of ImageNet detection, ImageNet localization, COCO detection, and COCO segmentation.

Deep learning is a form of machine learning that enables computers to learn from experience and understand the world in terms of a hierarchy of concepts. Because the computer gathers knowledge from experience, there is no need for a human computer operator to formally specify all the knowledge that the computer needs. The hierarchy of concepts allows the computer to learn complicated concepts by building them out of simpler ones; a graph of these hierarchies would be many layers deep. This book introduces a broad range of topics in deep learning. The text offers mathematical and conceptual background, covering relevant concepts in linear algebra, probability theory and information theory, numerical computation, and machine learning. It describes deep learning techniques used by practitioners in industry, including deep feedforward networks, regularization, optimization algorithms, convolutional networks, sequence modeling, and practical methodology; and it surveys such applications as natural language processing, speech recognition, computer vision, online recommendation systems, bioinformatics, and videogames. Finally, the book offers research perspectives, covering such theoretical topics as linear factor models, autoencoders, representation learning, structured probabilistic models, Monte Carlo methods, the partition function, approximate inference, and deep generative models. Deep Learning can be used by undergraduate or graduate students planning careers in either industry or research, and by software engineers who want to begin using deep learning in their products or platforms. A website offers supplementary material for both readers and instructors.

Deep Learning

At CHI2011 we ran a panel on how the CHI community handles the replicability of research and the reproducibility of findings. Careful scientific scholarship should build on firm foundations, which includes re-examining old evidences in the face of new findings. Yet, as a community that strives for novelty, we have very little motivation to look back and reconsider the validity of previous work. Thus, for CHI2013 we are planning a new venue, where replicated studies can be submitted, presented, and discussed. For CHI2012, we propose a SIG to discuss the preparations for how RepliCHI will work in its first year. We invite participation from those interested in setting an agenda for facilitating replication in HCI, including those who have begun using replication as a teaching method since RepliCHI at CHI2011.

RepliCHI SIG: from a panel to a new submission venue for replication

From smart homes that prepare coffee when we wake, to phones that know not to interrupt us during important conversations, our collective visions of HCI imagine a future in which computers understand a broad range of human behaviors. Today our systems fall short of these visions, however, because this range of behaviors is too large for designers or programmers to capture manually. In this paper, we instead demonstrate it is possible to mine a broad knowledge base of human behavior by analyzing more than one billion words of modern fiction. Our resulting knowledge base, Augur, trains vector models that can predict many thousands of user activities from surrounding objects in modern contexts: for example, whether a user may be eating food, meeting with a friend, or taking a selfie. Augur uses these predictions to identify actions that people commonly take on objects in the world and estimate a user's future activities given their current situation. We demonstrate Augur-powered, activity-based systems such as a phone that silences itself when the odds of you answering it are low, and a dynamic music player that adjusts to your present activity. A field deployment of an Augur-powered wearable camera resulted in 96% recall and 71% precision on its unsupervised predictions of common daily activities. A second evaluation where human judges rated the system's predictions over a broad set of input images found that 94% were rated sensible.

Augur: Mining Human Behaviors from Fiction to Power Interactive Systems

Social computing prototypes probe the social behaviors that may arise in an envisioned system design. This prototyping practice is currently limited to recruiting small groups of people. Unfortunately, many challenges do not arise until a system is populated at a larger scale. Can a designer understand how a social system might behave when populated, and make adjustments to the design before the system falls prey to such challenges? We introduce social simulacra, a prototyping technique that generates a breadth of realistic social interactions that may emerge when a social computing system is populated. Social simulacra take as input the designer’s description of a community’s design—goal, rules, and member personas—and produce as output an instance of that design with simulated behavior, including posts, replies, and anti-social behaviors. We demonstrate that social simulacra shift the behaviors that they generate appropriately in response to design changes, and that they enable exploration of “what if?” scenarios where community members or moderators intervene. To power social simulacra, we contribute techniques for prompting a large language model to generate thousands of distinct community members and their social interactions with each other; these techniques are enabled by the observation that large language models’ training data already includes a wide variety of positive and negative behavior on social media platforms. In evaluations, we show that participants are often unable to distinguish social simulacra from actual community behavior and that social computing designers successfully refine their social computing designs when using social simulacra.

/pdf/social-simulacra-creating-populated-prototypes-for-social-1kle76am.pdf

Social Simulacra: Creating Populated Prototypes for Social Computing Systems

In the physical classroom, peer interactions motivate students and expand their perspective. We suggest that synchronous peer interaction can benefit massive online courses as well. Talkabout organizes students into video discussion groups and allows instructors to determine group composition and discussion content. Using Talkabout, students pick a discussion time that suits their schedule. The system groups the students into small video discussions based on instructor preferences such as gender or geographic balance. To date, 2,474 students in five massive online courses have used Talkabout to discuss topics ranging from prejudice to organizational theory. Talkabout discussions are diverse: in one course, the median six-person discussion group had students from four different countries. Students enjoyed discussing in these diverse groups: the average student participated for 66 minutes, twice the course requirement. Students in more geographically distributed groups also scored higher on the final, suggesting that distributed discussions have educational value.

https://hci.stanford.edu/publications/2014/PeerStudio/LAS2014-CambreTalkabout.pdf

Talkabout: small-group discussions in massive global classes

Crowd-powered systems combine computation with human intelligence, drawn from large groups of people connecting and coordinating online. These hybrid systems enable applications and experiences that neither crowds nor computation could support alone. 
Unfortunately, crowd work is error-prone and slow, making it difficult to incorporate crowds as first-order building blocks in software systems. I introduce computational techniques that decompose complex tasks into simpler, verifiable steps to improve quality, and optimize work to return results in seconds. These techniques develop crowdsourcing as a platform so that it is reliable and responsive enough to be used in interactive systems. 
This thesis develops these ideas through a series of crowd-powered systems. The first, Soylent, is a word processor that uses paid micro-contributions to aid writing tasks such as text shortening and proofreading. Using Soylent is like having access to an entire editorial staff as you write. The second system, Adrenaline, is a camera that uses crowds to help amateur photographers capture the exact right moment for a photo. It finds the best smile and catches subjects in mid-air jumps, all in realtime. Moving beyond generic knowledge and paid crowds, I introduce techniques to motivate a social network that has specific expertise, and techniques to data mine crowd activity traces in support of a large number of uncommon user goals. 
These systems point to a future where social and crowd intelligence are central elements of interaction, software, and computation. (Copies available exclusively from MIT Libraries, Rm. 14-0551, Cambridge, MA 02139-4307. Ph. 617-253-5668; Fax 617-253-1690.)

Michael S. Bernstein

Papers

RepliCHI SIG: from a panel to a new submission venue for replication

Augur: Mining Human Behaviors from Fiction to Power Interactive Systems

Social Simulacra: Creating Populated Prototypes for Social Computing Systems

Talkabout: small-group discussions in massive global classes

Crowd-powered systems