Despite recent breakthroughs in the field of artificial intelligence (AI) -
or more specifically machine learning (ML) algorithms for object recognition
and natural language processing - it seems to be the majority view that current
AI approaches are still no real match for natural intelligence (NI). More
importantly, philosophers have collected a long catalogue of features which
imply that NI works differently from current AI not only in a gradual sense,
but in a more substantial way: NI is closely related to consciousness,
intentionality and experiential features like qualia (the subjective contents
of mental states) and allows for understanding (e.g., taking insight into
causal relationships instead of 'blindly' relying on correlations), as well as
aesthetical and ethical judgement beyond what we can put into (explicit or
data-induced implicit) rules to program machines with. Additionally,
Psychologists find NI to range from unconscious psychological processes to
focused information processing, and from embodied and implicit cognition to
'true' agency and creativity. NI thus seems to transcend any neurobiological
functionalism by operating on 'bits of meaning' instead of information in the
sense of data, quite unlike both the 'good old fashioned', symbolic AI of the
past, as well as the current wave of deep neural network based, 'sub-symbolic'
AI, which both share the idea of thinking as (only) information processing. In
the following I propose an alternative view of NI as information processing
plus 'bundle pushing', discuss an example which illustrates how bundle pushing
can cut information processing short, and suggest first ideas for scientific
experiments in neuro-biology and information theory as further investigations.

该论文提出了一种将自然智能视为信息处理与‘bundle pushing’相结合的替代观点，并探讨了一个例子，说明 bundle pushing 如何缩短信息处理，同时建议进行神经生物学和信息理论的科学实验进一步研究。

qualia 的目的：如果人类思考不仅仅是信息处理，会怎么样？

The purpose of qualia: What if human thinking is not (only) information processing?

Introduction: In contrast to current AI technology, natural intelligence --
the kind of autonomous intelligence that is realized in the brains of animals
and humans to attain in their natural environment goals defined by a repertoire
of innate behavioral schemata -- is far superior in terms of learning speed,
generalization capabilities, autonomy and creativity. How are these strengths,
by what means are ideas and imagination produced in natural neural networks?
Methods: Reviewing the literature, we put forward the argument that both our
natural environment and the brain are of low complexity, that is, require for
their generation very little information and are consequently both highly
structured. We further argue that the structures of brain and natural
environment are closely related.
Results: We propose that the structural regularity of the brain takes the
form of net fragments (self-organized network patterns) and that these serve as
the powerful inductive bias that enables the brain to learn quickly, generalize
from few examples and bridge the gap between abstractly defined general goals
and concrete situations.
Conclusions: Our results have important bearings on open problems in
artificial neural network research.

自然神经网络的结构规律采用了自组织网络模式，这种归纳偏见使其能够快速学习、从少量数据推广并填补抽象目标与具体情况之间的差距，为人工神经网络研究中的一些开放问题提供了重要的参考意见。

自然智能理论

A Theory of Natural Intelligence

Research on both natural intelligence (NI) and artificial intelligence (AI)
generally assumes that the future resembles the past: intelligent agents or
systems (what we call 'intelligence') observe and act on the world, then use
this experience to act on future experiences of the same kind. We call this
'retrospective learning'. For example, an intelligence may see a set of
pictures of objects, along with their names, and learn to name them. A
retrospective learning intelligence would merely be able to name more pictures
of the same objects. We argue that this is not what true intelligence is about.
In many real world problems, both NIs and AIs will have to learn for an
uncertain future. Both must update their internal models to be useful for
future tasks, such as naming fundamentally new objects and using these objects
effectively in a new context or to achieve previously unencountered goals. This
ability to learn for the future we call 'prospective learning'. We articulate
four relevant factors that jointly define prospective learning. Continual
learning enables intelligences to remember those aspects of the past which it
believes will be most useful in the future. Prospective constraints (including
biases and priors) facilitate the intelligence finding general solutions that
will be applicable to future problems. Curiosity motivates taking actions that
inform future decision making, including in previously unmet situations. Causal
estimation enables learning the structure of relations that guide choosing
actions for specific outcomes, even when the specific action-outcome
contingencies have never been observed before. We argue that a paradigm shift
from retrospective to prospective learning will enable the communities that
study intelligence to unite and overcome existing bottlenecks to more
effectively explain, augment, and engineer intelligences.

本文提出了一种更好的学习智能的方法 —— 前瞻性学习，指的是智能代理在学习未知的、前所未见的任务时，更新内部模型以适用于未来任务。这种学习方法依赖于四个关键因素：连续性学习、前瞻性约束、好奇心和因果估计。