IJCAI 2016 Workshop on
Deep Learning for Artificial Intelligence (DLAI)
New York Midtown Hilton Hotel | Room: Concourse A
10 July 2016 | New York, NY (USA)
Papers and Presentations |
Topics of Interest |
Paper Submissions |
Deep Learning (DL) techniques have ignited a firestorm of attention
from academia and industry due to their success on many benchmarks and
applications, such as those pertaining to image recognition, automated
caption generation, speech recognition, natural language processing,
and game AI. To date, most DL efforts have focused on analysis tasks
(e.g., classification, regression), and have not involved the use of
symbolic representations and inference methods commonly used in
AI. For example, few DL efforts have focused on synthesis tasks (e.g.,
planning, scheduling, and design), involved the use of logical or
graphical models, or made other tight connections with AI. Much could
be gained by investigating how best to integrate AI and DL techniques;
this would interest researchers studying a variety of topics.
The DLAI workshop will provide a forum for
describing and learning about promising methods that use DL techniques
in AI processes, highlighting initial contributions/benefits, and
encouraging further work, thereby fostering connections among two
somewhat disparate groups of researchers. We intend it to interest AI
researchers who investigate the use of DL techniques in their studies
and DL researchers who seek to understand/explore the roles that DL
techniques can perform in, for example, systems that can reason with
interpreted sensor data.
This workshop will include three primary components. The first is
invited talks by senior researchers who have investigated DLAI, can
provide visions for what these types of integrations can accomplish,
and can identify important open research issues. Second, we will host
presentations of accepted papers (please
see paper submissions for details). Finally, we
will include a panel of active DLAI researchers who will be asked to
discuss key issues that require attention, and speakers from industry
(or government) who have interest in solving problems using DLAI
The integration of AI and DL techniques could interest researchers
studying the following topic areas (among others):
We will host presentations by two invited speakers, listed (alphabetically) below.
- Caption generation
- Cognitive architectures
- Commonsense reasoning
- Episodic reasoning
- Intelligent agents (e.g., planning and acting, goal reasoning)
- Machine learning (e.g., deep, reinforcement, statistical relational, transfer)
- Model-based reasoning
- Narrative intelligence
- Temporal reasoning
- Visual reasoning
- Ambient intelligence
- Autonomic computing
- Biomedical systems
- Computer games
- Image processing (e.g., security/surveillance tasks)
- Information retrieval and reuse
- Intelligent tutoring systems
- Language translation
- Plan replay
- Reasoning from experiences mined from text, multimedia, or human interactions
- Recommender systems
- Robotics (e.g., control of autonomous vehicles)
- Text reuse
- User modeling
We welcome contributions that, for example, describe prior or ongoing
work on DLAI, describe key issues that require further research, or
highlight relevant challenges of interest to AI and DL researchers and
practitioners, and plans for addressing them. In particular, we
welcome four types of submissions:
- Pieter Abbeel (Associate Professor, EECS, UC
- Title: Deep Reinforcement Learning for Robotics
- Abstract: Deep learning has enabled significant advances in supervised
learning problems such as speech recognition and visual
recognition. Reinforcement learning (RL) provides only a weaker
supervisory signal, posing additional challenges in the form of
temporal credit assignment and exploration. Nevertheless, deep
RL has already enabled learning to play
Atari games from raw pixels (without access to the underlying
game state) and learning certain types of visuomotor
manipulation primitives. I will discuss major challenges for,
as well as some preliminary promising results towards, making
deep RL applicable to real robotic problems.
- BIO: Pieter Abbeel works in
machine learning and robotics, in particular his research is on
making robots learn from people (apprenticeship learning) and
how to make robots learn through their own trial and error
(RL). His robots have learned: advanced
helicopter aerobatics, knot-tying, basic assembly, and
organizing laundry. He has won various awards, including best
paper awards at ICML and ICRA, the Sloan Fellowship, the Air
Force Office of Scientific Research Young Investigator Program
(AFOSR-YIP) award, the Office of Naval Research Young
Investigator Program (ONR-YIP) award, the DARPA Young Faculty
Award (DARPA-YFA), the National Science Foundation Faculty
Early Career Development Program Award (NSF-CAREER), the
Presidential Early Career Award for Scientists and Engineers
(PECASE), the CRA-E Undergraduate Research Faculty Mentoring
Award, the MIT TR35, the IEEE Robotics and Automation Society
(RAS) Early Career Award, and the Dick Volz Best
U.S. Ph.D. Thesis in Robotics and Automation Award.
- David Gunning (DARPA Information Innovation Office (I2O))
- Title: Explainable AI
- Abstract: TBD
- BIO: Dr. Gunning is an experienced technology manager, with over 30
years experience in the research and development of advanced
software systems. He has participated in the conceptualization
and early prototype development of cutting-edge software, from
a variety of technical areas including artificial intelligence, machine
learning, data analytics, information systems, language
understanding, human-computer interaction, and cognitive
Self-contained submissions must be no longer than seven pages (i.e.,
six pages for the main text of the paper, and one additional page for
(only) references) in PDF format for letter-size (8.5 x 11)
paper. Please use
templates, and include author names, affiliations, and email
addresses on the first page. Submissions should be made through
- Theoretical and/or empirical analyses
- System demonstrations
- Planned research or application
- Position papers
Papers will be subject to peer review by the workshop program
committee. Selection criteria include originality of ideas,
correctness, clarity and significance of results, and quality of
The following is an incomplete list of related publications concerning
DLAI that go beyond standard classification and regression tasks
(e.g., object identification, action recognition, sentiment analysis).
We welcome additional recommendations for this list.
- 24 April 2016: Paper submissions due
- 18 May 2016: Reviews due to EasyChair
- 20 May 2016: Reviewing responses due to authors
- 3 June 2016: Camera-ready deadline
- 10 July 2016: DLAI Workshop
While highly relevant, few (if any) DLAI publications related to
reinforcement learning (RL) are listed here because Deep RL is the focal topic of
workshop at IJCAI-16.
- Aditya, S., Yang, Y., Baral, C., Fermuller, C., & Aloimonos,
Y. (2015). Visual common-sense for scene understanding using
perception, semantic parsing and reasoning. In L. Morgenstern,
T. Patkos, & R. Sloan (Eds.) Logical Formalizations of
Commonsense Reasoning (Technical Report SS-15-04). Stanford,
CA: AAAI Press.
- Antol, S., Agrawal, A., Lu, J., Mitchell, M., Batra, D., Lawrence
Zitnick, C., & Parikh,
D. (2015). VQA: Visual question
answering. In Proceedings of the IEEE International
Conference on Computer Vision (pp. 2425-2433). Santiago,
Chile: IEEE Press.
- Doshi, J., Kira, Z., & Wagner, A. (2015). From deep learning to
episodic memories: Creating categories of visual experiences. In
Proceedings of the Third Annual Conference on Advances in Cognitive Systems.
Atlanta, GA: [http://www.cogsys.org/papers/ACS2015/article15.pdf].
- Krishna, R., Zhu, Y., Groth, O., Johnson, J., Hata, K., Kravitz,
J., Chen, S., Kalantidis, Y., Li, L.-J., Shamma, D.A., Bernstein,
M.S., & Li, F.-F. (2016). Visual genome: Connecting language and
vision using crowdsourced dense image annotations. arXiv preprint
- Maynord, M., Bhattacharya, S., & Aha, D.W. (2016). Image
surveillance assistant. In O. Bulan, R. Loce, & Y. Cai (Eds.)
Computer Vision Applications in Surveillance and Transportation:
Papers from the WACV-16 Workshop. Lake Placid, NY:
- Nath, A., & Domingos,
P. M. (2015). Learning relational sum-product
networks. Proceedings of the Twenty-Ninth AAAI Conference
on Artificial Intelligence (pp. 2878-2886). Austin, TX: AAAI
- Rocktäschel, T., Grefenstette, E., Hermann, K. M.,
Kočiský, T., & Blunsom, P. (2015). Reasoning about
entailment with neural attention. arXiv preprint [arXiv:1509.06664].
M. (2016). Writing stories with help from recurrent neural
networks. In Proceedings of the Thirtieth AAAI Conference
on Artificial Intelligence. Phoenix, AZ: AAAI Press.
- Tamar, A., Levine, S., & Abbeel, P. (2016). Value iteration networks.
arXiv preprint, [arXiv:1602.02867].
- Wang, H., & Yeung, D.-Y. (2016). Towards Bayesian deep learning: A survey. arXiv preprint, [arXiv:1604.01662v2].
- Yang, Y., Li, Y., Fermuller, C., & Aloimonos,
Y. (2015). Neural self talk: Image understanding via continuous
questioning and answering. arXiv preprint,
- Zhang, P., Goyal, Y., Summers-Stay, D., Batra, D., & Parikh,
D. (2015). Yin and yang: Balancing and answering binary visual
questions. arxiv preprint. [arXiv:1511.05099].
- Is it appropriate to submit papers on Deep Reinforcement Learning to
- Deep RL relates to the subject of this DLAI workshop, in that it
combines elements of DL and AI. However, there is a more specific
workshop at IJCAI-16, namely the Deep RL Workshop, that is a better
fit for Deep RL submissions. However, if a Deep RL researcher wished to
present at both workshops, or otherwise could not attend the
Deep RL workshop, please contact the Co-Chairs to discuss their situation.
- Are Deep Reinforcement Learning researchers welcome to attend this
- I can't register! What should I do?
- On 2016 June 27, paper presenter Federico Castenado notified us of this problem. We learned
that, indeed, the registration system is not showing W20 now. Instead, people who still wish
to register should instead register for W21, which will apply the same fees and provide access to W20,
and then email Francisco Cruz Mencía [tito (at) iiia.csic.es] who, at your request, will swap W21 for W20.
- How many people are registered?
- As of 2016 June 28, we have 113 registered and are expecting at least 2 more. (This is the first of
the 41 IJCAI-16 workshops to max out in capacity.) Carol Hamilton is assisting in searching for
a larger room. So...thanks to everyone who will give presentations for making this an attractive
event to attend!
- Will you hold any followup events?
- Given the turnout, we will consider this. We welcome recommendations, not only for venues
(e.g., IJCAI-17), but also for topics, invited speakers, and related suggestions. Please
send them to David Aha [david.aha (at) nrl.navy.mil].
- When and where should I set up my poster?
- We'll share the hotel's concourse level with only 1 other workshop. We've
plenty of space to have the posters outside our workshop room (Concourse A)
and can have them up during the afternoon coffee break, immediately following
the 3:00-3:30pm poster session. So! Please feel free to set them up whenever
you'd like on 10 July; we'll place them just outside of the workshop room, with
the goal of attracting even more attention to them throughout that time period.
- (2016 January 11): Web site is live
- (2016 March 28): Added a few links to related work
- (2016 April 1): Pieter Abbeel accepted an invitation to be an invited speaker
- (2016 April 12): Dave Gunning accepted an invitation to be an invited speaker
- (2016 April 15): Extended the submission deadline to 24 April 2016 (end of day)
- (2016 April 25): We have received 16 submissions; bidding has started; will delay response date
- (2016 April 27): Revised all dates; notified authors & PC
- (2016 May 23): Notifications sent to all authors
- (2016 May 31): Due date for camera-ready papers
- (2016 June 13): First draft of agenda uploaded
- (2016 June 28): We have 114 Registrants! We're asking for a larger room.
- (2016 June 30): Whew: We've been given a larger room! We'll be in Concourse A on the Concourse level.
- (2016 June 30): Papers are finally online. See here.
- (2016 July 6): Final agenda uploaded.