Lifelong Learning - learn as "humans do"
Accumulating knowledge learned in the past and
using it to learn more knowledge
"Lifelong Machine Learning."
by Z. Chen and B. Liu, Morgan & Claypool Publishers, November 2016.
Lifelong Machine Learning Tutorial. (Title: lifelong machine learning and computer reading the Web), KDD-2016, August 13-17, 2016, San Francisco, USA.
Lifelong Machine Learning Tutorial, IJCAI-2015, July 25-31, 2015, Buenos Aires, Argentina
Statistical learning algorithms like deep NN, SVM, HMM, CRF, and topic
modeling have been very successful in machine learning and data mining
applications. Given a dataset, such an algorithm simply runs on the
dataset to produce a model without considering any
related information or past learning results. Although these algorithms
can still be improved, such single task and isolated algorithmic
approaches to machine learning have their limits, e.g., it needs a large
number of training examples and is only suitable for well-defined and
narrow tasks. Looking ahead, the question is how to deal with
these limitations to improve machine learning further. I believe
the answer is lifelong machine learning or simply
lifelong learning, which tries to mimic "human learning" (I use quotes because we don't know how humans learn)).
The key characteristic of "human learning" is that humans
learn continuously - we retain the knowledge gained from the past learning and
use the knowledge to help future learning and problem solving. Existing
isolated machine learning algorithms are not capable of doing that.
However, without the lifelong
machine learning capability, AI systems and machine learning will
probably never be truly
intelligent. We believe that now is the right time to explore lifelong
learning. Big data offers a golden opportunity
for lifelong learning because its large volume and diversity (very
important) give us abundant information for discovering rich
and commonsense knowledge automatically, which can enable an intelligent
learning agent to perform continuous learning, to accumulate
knowledge learned previously, and to become more and more knowledgeable
and better and better at learning.
In my life, I don't remember that anyone has ever given
me 1000 positive and 1000 negative examples and asked me to
build a classifier of any kind. Since I have
accumulated so much knowledge in the past, I can learn mostly with little
effort and few examples. If I don't have the accumulated
knowledge, even if you give 2000 training examples, I will not be able to
learn. For example, I don't understand Arabic. If you give me 2000 Arabic
documents and ask me to build a classifier, I cannot do it (without using
Transfer learning, multitask learning, and lifelong learning
- Characterisitcs of lifelong learning: (1) learning continuously, (2) accumulating the previously learned knowledge, and (3) using the knowledge to learn more knowledge.
- Transfer learning vs. lifelong learning: Transfer learning
uses the source domain labeled data to help target domain learning.
Unlike lifelong learning, transfer learning is not
continuous and has no knowledge retention. The source must be similar
to the target (which are normally selected by the user).
It is also only one-directional: source helps target, but not
the other way around.
- Multitask learning vs. lifelong learning: Multitask learning
optimizes learning of multiple tasks. Although it is possible to make
it continuous, multitask learning does not retain any explicit knowledge
except data, and when the number of task is really large, it is hard to
re-learn everything when faced with a new task.
- Lifelong Unsupervised Learning (ICML-2014, KDD-2014, WWW-2016, AAAI-2016, EMNLP-2016): Our first work is Lifelong Topic Modeling (ICML-2014, KDD-2014, WWW-2016). The system retains and consolidates the knowledge or topics learned from previous tasks and uses the knowledge suitably for future modeling. This paradigm is powerful because different domains or tasks share a great deal of concepts or topics, which can be exploited to generate much better topics for the new task. The other two papers are about lifelong unsupervised information extraction.
- Lifelong Supervised Learning (ACL-2015, KDD-2016): The first work is about sentiment classification. The second work is about cumulative learning and self-motivated learning (or self-driven learning), which learns in the open world rather than the closed world as classific learning.
- Lei Shu, Bing Liu, Hu Xu, and Annice Kim. Lifelong-RL: Lifelong Relaxation Labeling for Separating Entities and Aspects in Opinion Targets. Proceedings of 2016 Conference on Empirical Methods in Natural Language Processing (EMNLP-2016), November 1–5, 2016, Austin, Texas, USA.
- Geli Fei, Shuai Wang, and Bing Liu. 2016. Learning Cumulatively to Become More Knowledgeable. Proceedings of SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD-2016), August 13-17, San Francisco, USA.
- Shuai Wang, Zhiyuan Chen, and Bing Liu. Mining Aspect-Speciﬁc Opinion using a Holistic Lifelong Topic Model. Proceedings of the International World Wide Web Conference (WWW-2016), April 11-15, 2016, Montreal, Canada.
- Qian Liu, Bing Liu, Yuanlin Zhang, Doo Soon Kim and Zhiqiang Gao. Improving Opinion Aspect Extraction using Semantic Similarity and Aspect Associations. Proceedings of Thirtieth AAAI Conference on Artificial Intelligence (AAAI-2016), February 12–17, 2016, Phoenix, Arizona, USA.
- Zhiyuan Chen, Nianzu Ma and Bing Liu. Lifelong Learning for Sentiment Classification. Proceedings of the 53st Annual Meeting of the Association for Computational Linguistics (ACL-2015, short paper), 26-31, July 2015, Beijing, China.
- Zhiyuan Chen and Bing Liu. Mining Topics in Documents: Standing on the Shoulders of Big Data.. Proceedings of the 20th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD 2014), August 24-27, New York, USA. [Code] [Dataset]
- Zhiyuan Chen and Bing Liu. Topic Modeling using Topics from Many Domains, Lifelong Learning and Big Data. Proceedings of the 31st International Conference on Machine Learning (ICML 2014), June 21-26, Beijing, China.
Created on Sep 24, 2014 by Bing Liu.