2208 11561 Deep Symbolic Learning: Discovering Symbols and Rules from Perceptions

Finally, in section 7, we have discussed our results and outlined the pros/cons of using hyperdimensional vectors to fuse learning systems together at a symbolic level as well as what future work is necessary. Not everyone agrees that neurosymbolic AI is the best way to more powerful artificial intelligence. symbolic machine learning Serre, of Brown, thinks this hybrid approach will be hard pressed to come close to the sophistication of abstract human reasoning. Our minds create abstract symbolic representations of objects such as spheres and cubes, for example, and do all kinds of visual and nonvisual reasoning using those symbols.

Our results confirm the notion that hyperdimensional representations can be useful in VSA and symbolic reasoning systems. It is also important to note that hyperdimensional vectors have not yet been effectively used to represent dense RGB images in prior work. This potentially opens up new avenues for combining symbolic reasoning and ML methods. Hyperdimensional representations produced by converting the output of deep hashing networks into symbolic inference structures allows the use of fuzzy logic systems, of which the use of HILs in our experiments are a simple example of. Since HIL structures can be fused across different modalities, this increases the robustness and interpretability of the inference process.

Artificial Intelligence in Physical Sciences: Symbolic Regression Trends and Perspectives

Nevertheless, big data would be hard to handle without the incorporation of AI, or, in other words, the assemblage of methods and skills that allow humans and machines to execute assignments that only intelligent entities can (e.g., perceive, reason, and act) [11]. Samuel’s Checker Program[1952] — Arthur Samuel’s goal was to explore to make a computer learn. The program improved as it played more and more games and ultimately defeated its own creator.

Neural AI is more data-driven and relies on statistical learning rather than explicit rules. We investigate an unconventional direction of research that aims at converting neural networks, a class of distributed, connectionist, sub-symbolic models into a symbolic level with the ultimate goal of achieving AI interpretability and safety. To that end, we propose Object-Oriented Deep Learning, a novel computational paradigm of deep learning that adopts interpretable “objects/symbols” as a basic representational atom instead of N-dimensional tensors (as in traditional “feature-oriented” deep learning). For visual processing, each “object/symbol” can explicitly package common properties of visual objects like its position, pose, scale, probability of being an object, pointers to parts, etc., providing a full spectrum of interpretable visual knowledge throughout all layers. It achieves a form of “symbolic disentanglement”, offering one solution to the important problem of disentangled representations and invariance.

From Philosophy to Thinking Machines

We first tested how well a hyperdimensional representation of a given hashing network’s output can work with a HIL. That is, does the inclusion of a HIL (and by extension, hyperdimensional representations) obfuscate the classification, thereby worsening performance, or does it perhaps improve the performance? However, the nature of HIL’s structure may enable a better memorization of training examples. We trained individual Hash Networks to perform image classification and then compared their performance with and without a HIL. Performance is tracked by comparing the F1 score for classification to the number of training iterations for the Hash network or epochs. We were also interested in how the HIL affects the F1 score as the Hamming Distance threshold for similarity increases.

No explicit series of actions is required, as is the case with imperative programming languages. Programs were themselves data structures that other programs could operate on, allowing the easy definition of higher-level languages. The key AI programming language in the US during the last symbolic AI boom period was LISP. LISP is the second oldest programming language after FORTRAN and was created in 1958 by John McCarthy.

Just like the communicative success, we see that it quickly increases and stabilizes at 19 concepts, which are all concepts present in the CLEVR dataset. We cut off these figures after 2,500 of the 10,000 interactions, since the metrics reached a stable level. In this experiment, the tutor can use up to four words to describe the topic object.

Further, our method allows easy generalization to new object attributes, compositions, language concepts, scenes and questions, and even new program domains. It also empowers applications including visual question answering and bidirectional image-text retrieval. An LNN consists of a neural network trained to perform symbolic reasoning tasks, such as logical inference, theorem proving, and planning, using a combination of differentiable logic gates and differentiable inference rules. These gates and rules are designed to mimic the operations performed by symbolic reasoning systems and are trained using gradient-based optimization techniques.

One promising approach towards this more general AI is in combining neural networks with symbolic AI. In our paper “Robust High-dimensional Memory-augmented Neural Networks” published in Nature Communications,1 we present a new idea linked to neuro-symbolic AI, based on vector-symbolic architectures. In summary, symbolic AI excels at human-understandable reasoning, while Neural Networks are better suited for handling large and complex data sets.

One of the keys to symbolic AI’s success is the way it functions within a rules-based environment. Typical AI models tend to drift from their original intent as new data influences changes in the algorithm. Scagliarini says the rules of symbolic AI resist drift, so models can be created much faster and with far less data to begin with, and then require less retraining once they enter production environments. The technology actually dates back to the 1950s, says expert.ai’s Luca Scagliarini, but was considered old-fashioned by the 1990s when demand for procedural knowledge of sensory and motor processes was all the rage. Now that AI is tasked with higher-order systems and data management, the capability to engage in logical thinking and knowledge representation is cool again. Maybe in the future, we’ll invent AI technologies that can both reason and learn.