FlowingFlowing
← Back to all development cases
This case was judged by an anonymized AI reviewer and has not been independently reviewed by people; inspect the draft, cards, and both outputs to assess it.
AutocompleteWith cards better207 / 280 · beu-b467372e2c26aae7

Predictive Coding: a Theoretical and Experimental Review

Psychology and cognitive neuroscience · 2107.12979v4

FLOWING EVIDENCE BENCHMARK

How do we compare writing with and without cards?

At the same writing position, the same model produces one first result with cards and one without. All 280 pairs enter the overall anonymized review, which assesses whether a decision not to continue fits the draft.

Overall: 138 with-card wins, 26 ties, 107 no-card wins, 5 pairs where neither was usable, and 4 inconclusive judgments. Both sides wrote text in 218 pairs; 48 had a deliberate stop and 14 included an invalid output.

SAME DRAFT · WITH OR WITHOUT CARDS

01 · FIXED WRITING POSITION

MANUSCRIPT

Same writing position ▌

Matched autocomplete at the same draft position

02 · TWO MATCHED VERSIONS

SHARED BY BOTH

Draft, writing task, and generation model

A · With cards

Relevant literature cards supplied

B · Without cards

No literature cards supplied

Generation model

Same model and version

A → first output

B → first output

Anonymized AI judge

First outputs are anonymized as X and Y

Continuation review: accuracy, fit to the writing task and usability
Returns: X preferred / tie / Y preferred / both unusable

Review again with the answer order swapped

HOW DOES BLIND JUDGING WORK?

① Anonymize both outputs
The AI judge sees the same draft and both first outputs without knowing which received cards.

② Compare and swap order
The judge applies task-specific criteria in X/Y and then Y/X order.

③ Review disagreements
A third pass resolves disagreements. Incomplete reviews remain in the denominator.

How are positions and cards selected?

We selected 40 writing positions in each of seven fields, for 280 comparisons. The position and relevant literature cards were set before each comparison, so both versions began from the same draft.

The main result retains every frozen position, including places where cards might not add a useful next move. This avoids selecting cases based on their outputs, but 49% describes the whole cohort, not a win rate where evidence clearly had an opportunity to help. This cohort did not record that opportunity label for each position before generation, so those two groups cannot yet be separated reliably.

CARD USE · 140/280

Reported card use in continuations with text

The model reports whether a card affected its continuation. In 140 of 280 tests it produced continuation text and reported card use. The previous count of 150 included 10 cases without continuation text and has been corrected.

COMPARISON · 70 / 7 / 30

Which written version was better?

Of the 107 reported-use cases where both versions produced text, the blinded AI judge preferred the with-card version 70 times, found 7 about the same, and preferred the without-card version 30 times.

How are the two table percentages calculated?

ADOPTION COVERAGE

140reported use with text
÷
280all positions
=
50%

WITH-CARD VERSION PREFERRED

70with cards better
÷
107comparable tests
=
65%

The percentages use different groups: 50% counts continuations with text reporting card use among all 280 tests; 65% covers the 107 such cases with text on both sides. The generating model reports card use, while an anonymized AI judge compares writing quality.

Card use is self-reported by the generator. The anonymized AI judge sees the draft and source snippets, so source-aware preference remains possible; there was no independent human review. The overall result describes these 280 frozen positions, not a win probability for any writing task.

Manuscript writing position

Text excerpt · not a PDF page

Research manuscript · excerpt

Predictive Coding: a Theoretical and Experimental Review

5 Discussion and Future Directions

…ge-scale simulations of predictive coding networks with simultaneous learning has not yet been done. Finally, the implementation of precision in the brain remains largely mysterious. There are competing theories of precision being implemented through lateral connectivity ( K. Friston, (2005) ) , feedback superficial-to-superficial connectivity ( Shipp, (2016) ) , or primarily through subcortical processes and regions such as the pulvinar ( Kanai et al., (2015) ) . However, to our knowledge, there has been no real explicit experimental tests of the differing predictions made by these theories.

…orithm. Finally, while there has been considerable progress in predictive coding networks setup for the machine learning paradigm of hierarchical static image classification, there has been considerably little work experimenting with potentially more brain-like paradigm such as spatial predictive coding between pixels in an image, or temporal predictive coding over sequences. Indeed, there has not yet even been any large scale study of the performance characteristics of generalized coordinates on spatio-temporal video data, so precisely how effective they are in such settings remains unknown.

↓ Next paragraph was generated here

The highlighted text is near the writing position. This is a layout of the evaluation input, not a PDF screenshot.

A matched comparison at the same writing positionOverall AI review: With cards better

Both arms use the same model, task and manuscript context. Only the retrieved passages differ. The first recorded output from each arm is shown below.

Read overall AI review rationale

A = without cards · B = with cards

Both outputs provide valid continuations, but B fits substantially better. The draft ends by emphasizing gaps in large-scale simulations, precision implementation, and especially the lack of work on 'more brain-like paradigms' such as spatial/temporal predictive coding and generalized coordinates on video data. Output B directly addresses the relationship between predictive coding networks and backpropagation—a major machine learning paradigm gap that echoes the draft's concern about 'predictive coding networks setup for the machine learning paradigm.' The source material on inference learning approximating BP with unrealistic requirements is directly relevant and accurately summarized. Output A introduces Bayesian inference through sampling, which is a valid topic but represents a more abrupt thematic shift; the draft has not discussed sampling-based inference at all, whereas it has explicitly flagged the ML/backpropagation gap as an open problem. B maintains argumentative continuity by extending the 'machine learning paradigm' thread, while A opens a new theoretical dimension without grounding in the preceding text. B also better matches the sources provided, which emphasize the PCN/backpropagation relationship.

Evidence

Same model and position · retrieved passages supplied · first output

Preferred by blind review

Beyond these architectural and theoretical questions, a further open direction concerns the relationship between predictive coding networks and the backpropagation algorithm that dominates modern machine learning. Recent work has shown that inference learning can train predictive coding networks to approximate backpropagation, but this approximation carries unrealistic requirements, is only established for single-step weight updates, and remains computationally more costly than backpropagation itself.

Passages supplied to the Evidence version

Card 1 · Deep Predictive Coding Networks for Video Prediction and Unsupervised Learning

While great strides have been made in using deep learning algorithms to solve supervised learning tasks, the problem of unsupervised learning - leveraging unlabeled examples to learn about the structure of a domain - remains a difficult unsolved challenge. Here, we explore prediction of future frame…

Read full passage excerpt

While great strides have been made in using deep learning algorithms to solve supervised learning tasks, the problem of unsupervised learning - leveraging unlabeled examples to learn about the structure of a domain - remains a difficult unsolved challenge. Here, we explore prediction of future frames in a video sequence as an unsupervised learning rule for learning about the structure of the visual world. We describe a predictive neural network ("PredNet") architecture that is inspired by the concept of "predictive coding" from the neuroscience literature. These networks learn to predict future frames in a video sequence, with each layer in the network making local predictions and only forwarding deviations from those predictions to subsequent network layers. We show that these networks are able to robustly learn to predict the movement of synthetic (rendered) objects, and that in doing so, the networks learn internal representations that are useful for decoding latent object parameter …

Card 2 · PredNet and Predictive Coding: A Critical Review

PredNet, a deep predictive coding network developed by Lotter et al., combines a biologically inspired architecture based on the propagation of prediction error with self-supervised representation learning in video. While the architecture has drawn a lot of attention and various extensions of the mo…

Read full passage excerpt

PredNet, a deep predictive coding network developed by Lotter et al., combines a biologically inspired architecture based on the propagation of prediction error with self-supervised representation learning in video. While the architecture has drawn a lot of attention and various extensions of the model exist, there is a lack of a critical analysis. We fill in the gap by evaluating PredNet both as an implementation of the predictive coding theory and as a self-supervised video prediction model using a challenging video action classification dataset. We design an extended model to test if conditioning future frame predictions on the action class of the video improves the model performance. We show that PredNet does not yet completely follow the principles of predictive coding. The proposed top-down conditioning leads to a performance gain on synthetic data, but does not scale up to the more complex real-world action classification dataset. Our analysis is aimed at guiding future research …

Card 3 · Predictive Coding Can Do Exact Backpropagation on Convolutional and Recurrent Neural Networks

Predictive coding networks (PCNs) are an influential model for information processing in the brain. They have appealing theoretical interpretations and offer a single mechanism that accounts for diverse perceptual phenomena of the brain. On the other hand, backpropagation ( BP ) is commonly regarded…

Read full passage excerpt

Predictive coding networks (PCNs) are an influential model for information processing in the brain. They have appealing theoretical interpretations and offer a single mechanism that accounts for diverse perceptual phenomena of the brain. On the other hand, backpropagation ( BP ) is commonly regarded to be the most successful learning method in modern machine learning. Thus, it is exciting that recent work formulates inference learning ( IL ) that trains PCNs to approximate BP. However, there are several remaining critical issues: (i) IL is an approximation to BP with unrealistic/non-trivial requirements, (ii) IL approximates BP in single-step weight updates; whether it leads to the same point as BP after the weight updates are conducted for more steps is unknown, and (iii) IL is computationally significantly more costly than BP.

Card 4 · PreCNet: Next-Frame Video Prediction Based on Predictive Coding

Predictive coding, currently a highly influential theory in neuroscience, has not been widely adopted in machine learning yet. In this work, we transform the seminal model of Rao and Ballard (1999) into a modern deep learning framework while remaining maximally faithful to the original schema. The r…

Read full passage excerpt

Predictive coding, currently a highly influential theory in neuroscience, has not been widely adopted in machine learning yet. In this work, we transform the seminal model of Rao and Ballard (1999) into a modern deep learning framework while remaining maximally faithful to the original schema. The resulting network we propose (PreCNet) is tested on a widely used next-frame video prediction benchmark, which consists of images from an urban environment recorded from a car-mounted camera, and achieves state-of-the-art performance. Performance on all measures (MSE, PSNR, and SSIM) was further improved when a larger training set (2M images from BDD100k) pointed to the limitations of the KITTI training set. This work demonstrates that an architecture carefully based on a neuroscience model, without being explicitly tailored to the task at hand, can exhibit exceptional performance.

Card 5 · Deep Predictive Coding Networks

The quality of data representation in deep learning methods is directly related to the prior model imposed on the representations; however, generally used fixed priors are not capable of adjusting to the context in the data. To address this issue, we propose deep predictive coding networks, a hierar…

Read full passage excerpt

The quality of data representation in deep learning methods is directly related to the prior model imposed on the representations; however, generally used fixed priors are not capable of adjusting to the context in the data. To address this issue, we propose deep predictive coding networks, a hierarchical generative model that empirically alters priors on the latent representations in a dynamic and context-sensitive manner. This model captures the temporal dependencies in time-varying signals and uses top-down information to modulate the representation in lower layers. The centerpiece of our model is a novel procedure to infer sparse states of a dynamic model which is used for feature extraction. We also extend this feature extraction block to introduce a pooling function that captures locally invariant representations. When applied on a natural video data, we show that our method is able to learn high-level visual features. We also demonstrate the role of the top-down connections by s …

Model-reported card use

Card 3 · fact

Affected output: Recent work has shown that inference learning can train predictive coding networks to approximate backpropagation, but this approximation carries unrealistic requirements, is only established for single-step weight updates, and remains computationally more costly than backpropagation itself.

Model explanation: This row supplied the concrete limitations of the inference-learning approximation to backpropagation, which became the factual content of the new paragraph.

Card 4 · direction

Affected output: a further open direction concerns the relationship between predictive coding networks and the backpropagation algorithm that dominates modern machine learning

Model explanation: This row's framing of predictive coding's limited adoption in machine learning guided the choice to open a discussion direction on the predictive-coding/backpropagation relationship.

No Evidence

Same model and position · no retrieved passages · first output

Not preferred in this comparison

A further open question concerns the relationship between predictive coding and other accounts of cortical function, such as the hypothesis that the brain implements approximate Bayesian inference through sampling rather than through deterministic message passing.

Earlier card-use comparison

The earlier diagnostic compared only cases with reported card use and text on both sides; its scope differs from the overall AI review above.

The blinded model review preferred the Evidence version. Its original rationale appears below.

Read the original model judgment

Output A introduces an unsupported claim about 'approximate Bayesian inference through sampling rather than through deterministic message passing' — this specific alternative framework (sampling-based Bayesian inference) is not mentioned in any of the supplied sources. The sources discuss predictive coding, backpropagation, precision implementation, and video prediction, but not sampling-based Bayesian inference as a competing account of cortical function. Output B, by contrast, is directly grounded in the third source ('Predictive Coding Can Do Exact Backpropagation on Convolutional and Recurrent Neural Networks'), which explicitly states that inference learning trains PCNs to approximate BP with 'unrealistic/non-trivial requirements,' that it 'approximates BP in single-step weight updates; whether it leads to the same point as BP after the weight updates are conducted for more steps is unknown,' and that 'IL is computationally significantly more costly than BP.' Output B accurately paraphrases these three points. Both outputs are coherent continuations, but Output A contains material not found in the sources while Output B is fully source-grounded.

The passages are input material; their presence does not mean an output used them or that they support every claim. Draft and source passages are excerpted; no simulated scores or PDF appear here.