اشترك بالحزمة الذهبية واحصل على وصول غير محدود شمرا أكاديميا
تسجيل مستخدم جديدCommunicating Natural Programs to Humans and Machines
112
0
0.0
(
0
)
اسأل ChatGPT حول البحث
ﻻ يوجد ملخص باللغة العربية
The Abstraction and Reasoning Corpus (ARC) is a set of tasks that tests an agents ability to flexibly solve novel problems. While most ARC tasks are easy for humans, they are challenging for state-of-the-art AI. How do we build intelligent systems that can generalize to novel situations and understand human instructions in domains such as ARC? We posit that the answer may be found by studying how humans communicate to each other in solving these tasks. We present LARC, the Language-annotated ARC: a collection of natural language descriptions by a group of human participants, unfamiliar both with ARC and with each other, who instruct each other on how to solve ARC tasks. LARC contains successful instructions for 88% of the ARC tasks. We analyze the collected instructions as `natural programs, finding that most natural program concepts have analogies in typical computer programs. However, unlike how one precisely programs a computer, we find that humans both anticipate and exploit ambiguities to communicate effectively. We demonstrate that a state-of-the-art program synthesis technique, which leverages the additional language annotations, outperforms its language-free counterpart.
قيم البحث
اقرأ أيضاً
While machine learning algorithms excel at many challenging visual tasks, it is unclear that they can make predictions about commonplace real world physical events. Here, we present a visual and physical prediction benchmark that precisely measures t
his capability. In realistically simulating a wide variety of physical phenomena -- rigid and soft-body collisions, stable multi-object configurations, rolling and sliding, projectile motion -- our dataset presents a more comprehensive challenge than existing benchmarks. Moreover, we have collected human responses for our stimuli so that model predictions can be directly compared to human judgments. We compare an array of algorithms -- varying in their architecture, learning objective, input-output structure, and training data -- on their ability to make diverse physical predictions. We find that graph neural networks with access to the physical state best capture human behavior, whereas among models that receive only visual input, those with object-centric representations or pretraining do best but fall far short of human accuracy. This suggests that extracting physically meaningful representations of scenes is the main bottleneck to achieving human-like visual prediction. We thus demonstrate how our benchmark can identify areas for improvement and measure progress on this key aspect of physical understanding.
With the rise of machines to human-level performance in complex recognition tasks, a growing amount of work is directed towards comparing information processing in humans and machines. These studies are an exciting chance to learn about one system by
studying the other. Here, we propose ideas on how to design, conduct and interpret experiments such that they adequately support the investigation of mechanisms when comparing human and machine perception. We demonstrate and apply these ideas through three case studies. The first case study shows how human bias can affect how we interpret results, and that several analytic tools can help to overcome this human reference point. In the second case study, we highlight the difference between necessary and sufficient mechanisms in visual reasoning tasks. Thereby, we show that contrary to previous suggestions, feedback mechanisms might not be necessary for the tasks in question. The third case study highlights the importance of aligning experimental conditions. We find that a previously-observed difference in object recognition does not hold when adapting the experiment to make conditions more equitable between humans and machines. In presenting a checklist for comparative studies of visual reasoning in humans and machines, we hope to highlight how to overcome potential pitfalls in design or inference.
Distributed software is becoming more and more dynamic to support applications able to respond and adapt to the changes of their execution environment. For instance, service-oriented computing (SOC) envisages applications as services running over glo
bally available computational resources where discovery and binding between them is transparently performed by a middleware. Asynchronous Relational Networks (ARNs) is a well-known formal orchestration model, based on hypergraphs, for the description of service-oriented software artefacts. Choreography and orchestration are the two main design principles for the development of distributed software. In this work, we propose Communicating Relational Networks (CRNs), which is a variant of ARNs, but relies on choreographies for the characterisation of the communicational aspects of a software artefact, and for making their automated analysis more efficient.
Our goal is to learn a semantic parser that maps natural language utterances into executable programs when only indirect supervision is available: examples are labeled with the correct execution result, but not the program itself. Consequently, we mu
st search the space of programs for those that output the correct result, while not being misled by spurious programs: incorrect programs that coincidentally output the correct result. We connect two common learning paradigms, reinforcement learning (RL) and maximum marginal likelihood (MML), and then present a new learning algorithm that combines the strengths of both. The new algorithm guards against spurious programs by combining the systematic search traditionally employed in MML with the randomized exploration of RL, and by updating parameters such that probability is spread more evenly across consistent programs. We apply our learning algorithm to a new neural semantic parser and show significant gains over existing state-of-the-art results on a recent context-dependent semantic parsing task.
Can computers overcome human capabilities? This is a paradoxical and controversial question, particularly because there are many hidden assumptions. This article focuses on that issue putting on evidence some misconception related with future generat
ions of machines and the understanding of the brain. It will be discussed to what extent computers might reach human capabilities, and how it could be possible only if the computer is a conscious machine. However, it will be shown that if the computer is conscious, an interference process due to consciousness would affect the information processing of the system. Therefore, it might be possible to make conscious machines to overcome human capabilities, which will have limitations as well as humans. In other words, trying to overcome human capabilities with computers implies the paradoxical conclusion that a computer will never overcome human capabilities at all, or if the computer does, it should not be considered as a computer anymore.
الأسئلة المقترحة
سجل دخول لتتمكن من نشر تعليقات
التعليقات
جاري جلب التعليقات
سجل دخول لتتمكن من متابعة معايير البحث التي قمت باختيارها
