Can Large Language Models Understand Symbolic Graphics Programs?

https://www.arxiv.org/abs/2408.08313

by belter • 2 years ago

70 59 2 years ago

Lichtso2• 2 years ago

What an awful paper title, saying "Symbolic Graphics Programs" when they just mean "vector graphics". I don't understand why they can not just use the established term instead. Also, there is no "program" here, in the same way that coding HTML is not programming, as vector graphics are not supposed to be Turing complete. And where they pulled the "symbolic" from is completely beyond me.

tines2• 2 years ago

Also, there is no "program" here, in the same way that coding HTML is not programming, as vector graphics are not supposed to be Turing complete. And where they pulled the "symbolic" from is completely beyond me.

Aren't HTML and vector graphics descriptions both data structures that could be interpreted via a Turing-complete interpreter? I don't see the difference between HTML and a C AST in this regard.

jlarocco• 2 years ago

There's a slippery slope there. Is a Word document a program? Is a PNG file?

A computer program is a data structure, but data structures are not necessarily computer programs.

merlincorey5• 2 years ago

I'm more curious how they think LLM's can imagine things:

To understand symbolic programs, LLMs may need to possess the ability to imagine how the corresponding graphics content would look without directly accessing the rendered visual content

To my understanding, LLMs are predictive engines based upon their tokens and embeddings without any ability to "imagine" things.

As such, an LLM might be able to tell you that the following SVG is a black circle because it is in Mozilla documentation[0]:

    <svg viewBox="0 0 100 100" xmlns="http://www.w3.org/2000/svg">
      <circle cx="50" cy="50" r="50" />
    </svg>

However, I highly doubt any LLM could tell you the following is a "Hidden Mickey" or "Mickey Mouse Head Silhouette":

    <svg viewBox="0 0 175 175" xmlns="http://www.w3.org/2000/svg">
      <circle cx="100" cy="100" r="50" />
      <circle cx="50" cy="50" r="40" />
      <circle cx="150" cy="50" r="40" />
    </svg>

- [0] https://developer.mozilla.org/en-US/docs/Web/SVG/Element/cir...

randomdata• 2 years ago

without any ability to "imagine" things.

What's imagining, then? The way LLMs explore different predictive branches in order to find an optimal solution doesn't seem all that different than what I consider imagining: Thinking about what could be and considering different variations on that idea.

An LLM isn't a brain, so there is no implication of it being said in the truest human sense, but it seems like a decent analogy to me.

rel_ic• 2 years ago

Check out Act One of this This American Life episode https://www.thisamericanlife.org/803/transcript

TLDR: it seems like an LLM might be able to tell you your SVG is a "Mickey Mouse Head Silhouette"

montebicyclelo2• 2 years ago

Chat GPT:

Given the arrangement of three overlapping circles, it resembles the classic depiction of a *Mickey Mouse* head silhouette:
The two smaller circles represent Mickey's ears.
The larger circle represents his head.
This is a stylized version of the iconic Mickey Mouse logo.

Imo: In order to predict the next token for non-trivial tokens, (of which there are many on the training data), you do have to do some more complex thinking/reasoning than just a lookup of past training data.

jaymzcampbell3• 2 years ago

GPT4o certainly gave it a good go (prompt: "What cartoon character would you say this SVG code renders most closely to?"). That's maybe a little too "prompting" tbf:

""" The SVG code you provided creates an image with three circles:

A large circle at the center (at coordinates 100, 100) with a radius of 50. Two smaller circles at the top left (at coordinates 50, 50) and top right (at coordinates 150, 50) with a radius of 40 each. This configuration closely resembles the head and ears of Mickey Mouse. The larger circle represents the head, and the two smaller circles represent the ears. """

Slightly more generic prompt: "What is this code drawing? Can you suggest what it might be representing?"

""" What It Might Represent: Mickey Mouse Head: The arrangement resembles a simplified "Mickey Mouse" head, where the large central circle represents the head, and the two smaller circles represent the ears.

Abstract Art: Alternatively, it could be an abstract design with three overlapping circles of different sizes, suggesting some form of simple geometric art.

The circles are positioned such that the smaller ones are symmetrically placed above the larger central one, which is a typical representation of a head with ears. """

And more generic "I've been sent this code, what is it drawing?" ends with:

""" You will see three overlapping circles. The largest circle is near the center, while two smaller circles are positioned above and slightly to the left and right of the center. The arrangement may resemble a basic representation of a Mickey Mouse head shape, with the large circle as the face and the two smaller circles as ears. """

westurner2• 2 years ago

If teh LLM saves the SVG vector graphic to a raster image like a PNG and prompts with that instead, it will have no trouble labeling what's depicted in the SVG.

So, the task is "describe what an SVG depicts without saving it to a raster image and prompting with that"?

ctoth• 2 years ago

However, I highly doubt

I never understand why people don't just test this?

Claude 3.5 sonnet, first test, prompt: Any idea what this SVG is describing?:

This SVG describes a simple geometric composition consisting of three circles: 1. The largest circle is centered at coordinates (100, 100) with a radius of 50 units. 2. A smaller circle is positioned at the top-left, centered at (50, 50) with a radius of 40 units. 3. Another smaller circle is positioned at the top-right, centered at (150, 50), also with a radius of 40 units. The overall effect, if rendered, would resemble a basic face-like structure or a simplified Mickey Mouse silhouette, with two smaller circles (potentially representing ears or eyes) above a larger central circle (potentially representing a face or head). The SVG uses a viewBox of "0 0 175 175", which defines the coordinate system and aspect ratio for the image. Would you like me to create a visual representation of this SVG for you?

jchw• 2 years ago

Also, there is no "program" here, in the same way that coding HTML is not programming, as vector graphics are not supposed to be Turing complete.

I think the reason why we don't view HTML as a programming language is because it is explicitly designed to be a markup language that declares content rather than a series of instructions that is interpreted as a program. A program needn't demonstrate turing completeness to be a "computer program", it just needs to be a sequence of instructions that a computer executes. To me, that suggests that there's a degree of abstractness and subjectivity involved. For example, any SVG document could also be rewritten 1:1 with no loss in fidelity as a series of commands that has the same effect, as can pretty much any declarative markup language; what is actually happening during parsing is hard to distinguish from an interpreter. Humans can "know it when they see it", but I doubt there's an exact criteria that can go along with the human "feel" of what makes a program, a program.

brap• 2 years ago

Crazy that LLMs are able to do that when they’re not even trained on visual data.

dmd• 2 years ago

But... they are? What are you saying here?

katzinsky• 2 years ago

I think he probably means that the LLMs never see the rasterized image output.

skybrian• 2 years ago

It’s difficult to tell what they’re proposing from the title or abstract, but asking questions about an SVG image, given only the source code, seems like an interesting challenge.

There are a lot of ways to generate the same image in different ways, or to slightly change the source code, resulting in a very different image. This could be used to train a deeper understanding of program-like text.

slashdave12• 2 years ago

In the title and abstract there is a proposal that LLMs "understand", "reason", and "imagine." It's off putting to me, since LLMs are not capable of any of these.

mcbuilder• 2 years ago

Not in the traditional sense, but how are we to possibly gauge an LLMs "intelligence" without framing questions like this?

Kuinox7• 2 years ago

Can you come up with a proof with your bold statements, since currently leading researcher on the subject have no idea if that's the case ?

slashdave12• 2 years ago

Hard to prove a negative. On the other hand, relying on a poorly defined idea of "emergent behavior" when there is no architectural rational is rather poor form IMO. It's more wordy but safer to qualify these terms appropriately. That is, to test the "appearance" or "simulation" of human-like abilities.

stevenhuang6• 2 years ago

What a sorry world it would be if we avoided studying anything that someone found "off-putting".

Science seeks truth and challenges perspectives. It's not about studying what comforts you.

Some people just aren't cut out for the sciences.

abernard1• 2 years ago

Since people are not reading the paper, here were the conclusions:

Is it really easy to answer semantic reasoning questions over symbolic graphics programs? We provide an intriguing experiment to demonstrate that SVG programs can be quite difficult for LLMs to understand such that even if the corresponding rendered images are fairly easy for humans to recognize, all these powerful LLMs still fail dramatically, only reaching a chance-level accuracy.

They then fine tune the model for this problem, and it achieves better results.

magicalhippo• 2 years ago

I was recently playing with some LLMs regarding a relay driver circuit. ChatGPT generated a circuit diagram for me in ASCII art style, but it contained some errors. For fun I tried to ask some others if they could identify the errors.

Here's the diagram (formatting errors were from ChatGPT):

         +5V
           |
           |
          ( ) Relay Coil
           |
           +----->|----+
           |      Diode |
           |            |
           |            |
           +------------+
           |
           |
         Drain (D) of MOSFET
            |
          Source (S)
            |
          -----
           GND
            |
         Gate (G)---[220Ω]--- GPIO (3.3V MCU)
            |
          [10kΩ]
            |
           GND

Here's what Llama 3.1 70B said:

After reviewing the diagram, I've found a few potential issues:

- The diode is not connected to the relay coil in a way that would prevent back-EMF from damaging the MOSFET. Typically, the diode is connected in parallel with the relay coil, with the anode connected to the positive terminal of the coil and the cathode connected to the negative terminal. In this diagram, the diode appears to be connected in series with the relay coil, which is not correct.

- The MOSFET is not properly biased. The gate of the MOSFET is connected to the GPIO pin through a 220Ω resistor, but there is no pull-down resistor to ensure the gate is at a known state when the GPIO pin is not driving it. The 10kΩ resistor connected to GND is likely intended to be a pull-down resistor, but it's not clear if it's connected to the gate or not.

- The relay coil is not specified, but it's likely that it requires a higher current than the MOSFET can handle. Typically, a relay driver circuit would include a current-limiting resistor or a transistor to handle the high current required by the relay coil.

So it correctly identified one of the main issues, the flyback diode is wrong, but failed to correctly identify the issue with the MOSFET gate being directly tied to ground. I admit, it's not a typical way to draw things so fair enough, after all it did mention there was something fishy with the pull-down.

Claude 3 Haiku provided a very similar answer. It was a bit more explicit and included more details. For instance it thought there might be an issue driving the MOSFET from 3.3V, though failed to suggest a logic-level MOSFET, instead suggesting a level-shifter. It also suggested the 220 Ohm gate resistor might not be enough to protect the microcontroller. Though it also tried to correct the diagram but failed: it changed the diode to ->|<-, and didn't put it in parallel with the coil.

All in all, somewhat impressed.

However they get easily confused. I tried to improve the diagram, and asked again. Here's the updated diagram:

         +5V
           |
           |          
           |
           +----->|--------+
           |      Diode    |
           |               |
          ( ) Relay Coil   |
           |               |
           +---------------+
           |
           |
    Drain (D) of MOSFET
           |
           |]- Gate (G)-----+-[220Ω]--- GPIO (3.3V MCU)
           |                |
           |             [10kΩ]
       Source (S)           |
           |                |
           +----------------+
           |
         -----
          GND

And Llama 3.1 now says:

After examining the diagram, I notice a few potential issues:

- The diode is not connected to the relay coil in a way that would prevent back-EMF damage to the MOSFET. Typically, a flyback diode is connected in parallel with the relay coil, with the anode connected to the coil's positive terminal and the cathode connected to the coil's negative terminal. In this diagram, the diode appears to be connected in series with the relay coil, which is not correct. - The MOSFET's source (S) is not connected to GND. In a typical relay driver circuit, the source of the MOSFET is connected to GND to provide a return path for the current. - The 10kΩ resistor is not connected to anything. It's possible that it's intended to be a pull-down resistor for the MOSFET's gate, but it's not clear where it's supposed to be connected.

So it totally fails to "see" that the 10k is now correctly connected to ground, and that it shares the ground point (in the diagram) with the source of the MOSFET. At least it figures it's meant to be a pull-down.

However it did notice the diode is now correctly placed in parallel with the coil, and did again correctly see the diode was the wrong way around. So that's something.

edit: I changed the diode direction and asked if that fixed it. Llama 3.1 did not "see" that and suggested a change.

Claude 3 did much better and noticed the direction change and was happy.

As I said, was just fooling around with this and thought I'd share. They're confident, but sometimes oh-so-wrong. Still, using a service where you can ask the same question to multiple models is nice. That way it can be easier to see when they're confabulating.

Can Large Language Models Understand Symbolic Graphics Programs?

Related Stories