=Paper=
{{Paper
|id=Vol-1802/paper5
|storemode=property
|title=Reasoning with Deep Learning:
an Open Challenge
|pdfUrl=https://ceur-ws.org/Vol-1802/paper5.pdf
|volume=Vol-1802
|authors=Marco Lippi
|dblpUrl=https://dblp.org/rec/conf/aiia/Lippi16
}}
==Reasoning with Deep Learning:
an Open Challenge==
https://ceur-ws.org/Vol-1802/paper5.pdf
Reasoning with Deep Learning:
an Open Challenge
Marco Lippi
DISMI – Università degli Studi di Modena e Reggio Emilia
marco.lippi@unimore.it
Abstract. Building machines capable of performing automated reason-
ing is one of the most complex but fascinating challenges in AI. In partic-
ular, providing an effective integration of learning and reasoning mech-
anisms is a long-standing research problem at the intersection of many
different areas, such as machine learning, cognitive neuroscience, psy-
chology, linguistic, and logic. The recent breakthrough achieved by deep
learning methods in a variety of AI-related domains has opened novel
research lines attempting to solve this complex and challenging task.
1 Motivation
In the last decade, deep learning has brought a real revolution in the area of artifi-
cial intelligence (AI) and in many of its related fields, producing stunning results
in a variety of different application domains. In computer vision, image classifica-
tion and object detection systems can now be trained to recognize thousands of
different semantic categories [1], that sometimes are difficult to distinguish even
for humans. Speech recognition and music retrieval can be performed with an
accuracy that was hard to imagine only one decade ago [2]. For natural language
processing and understanding, tasks such as machine translation or sentiment
analysis have moved huge steps forward with respect to earlier state-of-the-art
systems [3]. In addition, in many of such contexts, these successful applications
have produced a tremendous impact also from a technological point of view, with
all major ICT companies in the world (Google, Facebook, Microsoft, IBM, etc.)
now actively working in the field of AI more than ever before, aiming to con-
tinuously develop more efficient and more accurate systems. Whereas these are
indeed impressive advancements, there is no doubt that many of the problems
that are really at the core of AI are far from being solved. This is particularly true
for those tasks that have to deal with reasoning operations, such as induction,
deduction, abduction, probabilistic inference, spatial or temporal reasoning, and
especially combinations of those. We can now build machines that can easily
and accurately translate a text between languages, that can spot whether an
object appears in an image or in a video, that are capable of recognizing spoken
language at very high accuracy levels, but which cannot yet answer higher-level
questions related to the content they have just processed. Building a machine
that can read any kind of short story, or watch a movie of any genre, and that
Marco Lippi
can answer simple questions about the plot and the characters, questions that a
child would certainly be able to answer, still remains a dream. Clearly, these are
extremely complex tasks, that humans learn to perform during the first years
of life, and that involve learning to analyze large amounts of information, to
extract and somehow store some form of knowledge from such information, and
finally to digest this information and reason about it.
2 Methods
Historically, there has always been a dichotomy between symbolic and sub-
symbolic (often named connectionist) frameworks to model reasoning [4]. The
symbolic approach has its roots in the study of logic and philosophy, and it
sees reasoning as the capability of deriving additional information from that al-
ready encoded in a collection of given symbols, by performing elaboration and
manipulation on the given structured representations. From the perspective of
connectionism, reasoning is instead the result or derivation of multiple, intercon-
nected, simple processing devices, one major example being neural networks. The
main motivation behind connectionism comes from cognitive neuroscience, since
the human neural circuitry is clearly capable of storing and retrieving knowl-
edge organized in short- and long-term memory, by continuously analyzing and
processing new, complex information, and reasoning upon it.
2.1 Pioneering approaches
Throughout the years, there have been many attempts to combine learning and
reasoning processes by integrating connectionist and symbolic paradigms. Be-
tween the 80s and the 90s, a significant number of pioneering works started to
circulate, such as connectionist approaches to encode semantic networks [5], or
knowledge-based artificial neural networks, named KBANNs [6]. Within this con-
text, research has been mainly directed along distinct but strongly intertwined
directions: (i) inserting background knowledge into the structure of neural net-
works, (ii) refining sets of rules via neural networks, (iii) extracting rules or
classification patterns from trained neural networks.
The main idea behind KBANNs is that of considering input-to-output paths
in a neural network as sub-symbolic realizations of some symbolic rules given in
advance: output units can be thought of as the final conclusions of the rules, input
units are supporting facts and hidden units represent intermediate conclusions.
Standard backpropagation can be applied to tune the weights of the network, by
employing a training set, as for standard neural networks. This framework can be
adopted both for initializing the structure of a neural network with background
knowledge, and to extract a set of refined rules from the final learned network,
thus addressing the long-standing problem of neural network interpretability.
Similar approaches had been proposed for Recurrent Neural Networks to handle
sequential data [7]. Despite having shown promising results in computational
biology tasks [6], KBANNs have found applications only in small-sized domains,
Reasoning with Deep Learning: an Open Challenge
and encoding simple rules. One of the main limitations of this model was in fact
due to the difficulty, in the 90s, of training deep neural networks, whose structure
was induced by complex rules. In this direction, the recent advancements of deep
learning could certainly offer a valuable contribution.
2.2 Combining symbolic and sub-symbolic methods
More recent attempts to combine symbolic and sub-symbolic techniques for rea-
soning include the research lines carried out by the so-called neural-symbolic
community [8]. Several theoretical results have been succesfully achieved in this
area. Many studies have been conducted on the neural binding problem [9], that
aims to explain how connections between different brain regions are coordinated,
so as to retrieve and manipulate information, activate distant neural circuits, and
finally perform reasoning. Other research has focused on the analysis of the ca-
pability of neural networks to represent modal and temporal logics [10] as well as
fragments of first-order logic [11, 12]. Despite being successfully applied in some
proof-of-concept settings, the existing neural-symbolic approaches still lack a
thorough application on large-scale, real-world problems.
Starting from a slightly different perspective, the area of statistical relational
learning [13] (also known as probabilistic inductive logic programming) was born
at the end of the 90s with similar goals. Statistical relational learning aims to
combine the expressive power of logic representations with models handling un-
certainty in data, such as statistical learning approaches and graphical models.
Few attempts have been made in the direction of employing neural networks
within the context. An example is given by ground-specific Markov logic net-
works [14], that allow to embed neural networks within the Markov logic frame-
work, by learning the weights of the probabilistic logic clauses. The method
has been successfully applied to bioinformatics and time-series forecasting, for
problems where there is a crucial need to model background knowledge, handle
structured data, and perform probabilistic inference. Yet, it was never used to
handle reasoning tasks.
2.3 Recent advances: deep learning
In the last years, the task of reasoning with (deep) connectionist models has
captured an enormous interest, that is evidenced by the approaches that have
been proposed by some of the big companies that are currently investing in deep
learning. This is the case of Neural Turing Machines by Google DeepMind [15],
Memory Networks developed at Facebook AI Research [16], Dynamic Memory
Networks proposed by MetaMind [17], the Neural Reasoner [18] by Huawei Tech-
nologies, and the Watson system developed by IBM [19]. Additional methods
that are worth mentioning in this context are Wolfram Alpha, a computational
knowledge engine which is capable of handling and manipulating encyclopedic
knowledge to perform question-answering, and the GeoS system developed by
the Allen Institute [20] which can solve geometry Scholastic Aptitude Tests at
the level of the average US students.
Marco Lippi
Many of such methods employ a purely sub-symbolic framework, relying on
supervised datasets to train a deep architecture from collections of examples.
Most of these approaches share the common idea that a connectionist model
aiming to perform reasoning has to maintain some memory that has to be effi-
ciently organized and queried in order to retrieve the information necessary to
provide solutions for the desired tasks. Memory Networks, for example, use a
dedicated neural network for each step in the process of retrieving the correct
answer to a given question: (i) computing feature representations for the input,
(ii) updating memory, (iii) combining input and memory to compute the output,
(iv) translate the output into an interpretable answer. In their original imple-
mentation, such model is presented in a purely supervised fashion, but extensions
to semi-supervised settings are considered as well.
3 Discussion
Although producing remarkable advancements, recent approaches to reasoning
with deep networks do not properly address the task of symbolic reasoning,
thus leaving the problem of neural network interpretability unsolved. Most of
the effort is in fact demanded to an efficient management of the memory of
the network, and to fast matching and retrieval algorithms. Some of the ex-
isting approaches have been compared on a collection of benchmarks, called
bAbI tasks [21], developed at Facebook AI Research. Such tasks include sim-
ple question answering problems, that typically require to perform some kind of
reasoning and answer with a single word. The following is an example:
In the afternoon Julie went to the park. Yesterday Julie was at school. Julie
went to the cinema this evening. Where did Julie go after the park ? Cinema
To answer such questions, the system needs to perform many, advanced op-
erations. First, it has to process the text and store the information in some form
of memory, since even a short story like the one in the above example contains
plenty of information. Then, it has to understand which are pieces of knowledge
that are relevant to a given question, in order to finally formulate some hypoth-
esis and provide the correct answer. These final steps include complex reasoning
mechanisms, such as deductions and uncertainty handling, as well as temporal
reasoning. Such skills are completely different from the technology that is present
in existing sophisticated question answering systems, that mainly exploit ency-
clopedic background knowledge and answer highly specific questions.
Big data. The recent, impressive success of deep learning across several, differ-
ent areas of AI is certainly strongly related to the availability of huge datasets,
that nowadays can be easily collected from various and heterogeneous data
sources over the Web, and also to the advancements in computer hardware per-
formance, that have dramatically reduced computational requirements. From
a theoretical point of view, models that are currently employed in many sys-
tems were already known decades ago, but efficient techniques for training them
Reasoning with Deep Learning: an Open Challenge
successfully have been proposed only in recent years [22]. This is the case, for
example, of Convolutional and Recurrent Neural Networks, now representing the
state-of-the-art in a wide number of tasks. The injection of background knowl-
edge in the structure of such networks is yet to be investigated.
Unsupervised learning. Among the open challenges, a crucial point is to au-
tomatically extract knowledge from data, and to encode it into a neural network
model, rather than employing expert-given knowledge. Clearly, most of the ex-
isting methods for information extraction and knowledge representation employ
supervised or at least semi-supervised data. But, in the future we expect that a
key contribution will come from unsupervised learning approaches, also to ex-
tract commonsense knowledge. The advantages of using unsupervised data are
undeniable: generating labeled corpora is in fact an extremely complex, time-
consuming and costly operation, whereas unsupervised data are everywhere,
available in a variety of different domains (text, video, audio, etc.). Unsuper-
vised learning algorithms could be employed to extract relevant features and
patterns from data. Although some algorithms for unsupervised learning have
played a crucial role for the development of the whole deep learning area, it is
widely recognized that a proper use of unsupervised data is still missing [22].
Incremental learning. Humans naturally implement a lifelong learning scheme,
continuously acquiring knowledge. Such a feature seems to be a crucial element
for the development of reasoning skills and thus it is likely that future attempts
to this task will need to implement a dynamic, on-line mechanism that incre-
mentally acquires knowledge, possibly by also changing the network topology.
Beyond the Turing test ? Reasoning tasks could certainly be employed in an
advanced version of the Turing test. Recently, the computer vision community
has proposed the Visual Turing Challenge [25] where automated vision systems
have to answer questions regarding the content of some images or videos, thus
requiring both visual and linguistic skills. Also the bAbI tasks [21] already men-
tioned represent another example of benchmark that in future could be inte-
grated with an advanced Turing test.
References
1. Krizhevsky, A., Sutskever, I., Hinton, G.E.: Imagenet classification with deep
convolutional neural networks. In: Advances in NIPS. (2012) 1097–1105
2. Lee, H., Pham., P.T., Largman, Y., Ng, A.Y.: Unsupervised feature learning for
audio classification using convolutional deep belief networks. In: Advances in NIPS.
(2009) 1096–1104
3. Cho, K., van Merrienboer B., Gülcehre, C., Bahdanau, D., Bougares, F., Schwenk,
H., Bengio, Y.: Learning phrase representations using rnn encoder-decoder for
statistical machine translation. In: Proceedings of EMNLP. (2014) 1724–1734
Marco Lippi
4. Dinsmore, J.: The symbolic and connectionist paradigms: closing the gap.
Lawrence Erlbaum (2014)
5. Shastri, L.: A connectionist approach to knowledge representation and limited
inference. Cognitive Science 12 (1988) 331–392
6. Towell, G.G., Shavlik, J.W.: Knowledge-based artificial neural networks. Artificial
intelligence 70 (1994) 119–165
7. Frasconi, P., Gori, M., Maggini, M., Soda, G.: Unified integration of explicit knowl-
edge and learning by example in recurrent networks. IEEE Transactions on Knowl-
edge and Data Engineering 7 (1995) 340–346
8. d’Avila Garcez, A., Gori, M., Hitzler, P., Lamb, L.C.: Neural-symbolic learning
and reasoning (dagstuhl seminar 14381). Dagstuhl Reports 4 (2015)
9. Feldman, J.: The neural binding problem (s). Cognitive neurodynamics 7 (2013)
1–11
10. Garcez, A.S.d., Lamb, L.C.: A connectionist computational model for epistemic
and temporal reasoning. Neural Computation 18 (2006) 1711–1738
11. Bader, S., Hitzler, P., Hölldobler, S.: Connectionist model generation: A first-order
approach. Neurocomputing 71 (2008) 2420–2432
12. Garcez, A.S.d., Lamb, L.C., Gabbay, D.M.: Neural-symbolic cognitive reasoning.
Springer Science & Business Media (2008)
13. Getoor, L., Taskar, B.: Introduction to Statistical Relational Learning (Adaptive
Computation and Machine Learning). The MIT Press (2007)
14. Lippi, M., Frasconi, P.: Prediction of protein β-residue contacts by markov logic
networks with grounding-specific weights. Bioinformatics 25 (2009) 2326–2333
15. Graves, A., Wayne, G., Danihelka, I.: Neural Turing machines. arXiv preprint
arXiv:1410.5401 (2014)
16. Sukhbaatar, S., Weston, J., Fergus, R., et al.: End-to-end memory networks. In:
Advances in neural information processing systems. (2015) 2440–2448
17. Socher, R., Chen, D., Manning, C.D., Ng, A.: Reasoning with neural tensor net-
works for knowledge base completion. In: NIPS. (2013) 926–934
18. Peng, B., Lu, Z., Li, H., Wong, K.F.: Towards neural network-based reasoning.
arXiv preprint arXiv:1508.05508 (2015)
19. Gliozzo, A., Biran, O., Patwardhan, S., McKeown, K.: Semantic technologies in
ibm watsontm. ACL 2013 (2013) 85
20. Seo, M., Hajishirzi, H., Farhadi, A., Etzioni, O., Malcolm, C.: Solving geometry
problems: Combining text and diagram interpretation. In: Proceedings of EMNLP.
(2015) 17–21
21. Weston, J., Bordes, A., Chopra, S., Rush, A.M., van Merriënboer, B., Joulin, A.,
Mikolov, T.: Towards ai-complete question answering: A set of prerequisite toy
tasks. arXiv preprint arXiv:1502.05698 (2015)
22. LeCun, Y., Bengio, Y., Hinton, G.: Deep learning. Nature 521 (2015) 436–444
23. Serre, T., Kouh, M., Cadieu, C., Knoblich, U., Kreiman, G., Poggio, T.: A theory
of object recognition: computations and circuits in the feedforward path of the
ventral stream in primate visual cortex. Technical report, MIT (2005)
24. Krawczyk, D.C., McClelland, M.M., Donovan, C.M.: A hierarchy for relational
reasoning in the prefrontal cortex. Cortex 47 (2011) 588–597
25. Shan, Q., Adams, R., Curless, B., Furukawa, Y., Seitz, S.M.: The visual turing
test for scene reconstruction. In: 2013 International Conference on 3D Vision-3DV
2013, IEEE (2013) 25–32