Papers › Contrastive Training of Complex-Valued Autoencoders for Object Discovery

Contrastive Training of Complex-Valued Autoencoders for Object Discovery

24 May 2023NeurIPS 2023 11arXiv:2305.15001archive 2025-07-28

Aleksandar Stanić, Anand Gopalakrishnan, Kazuki Irie, Jürgen Schmidhuber

Current state-of-the-art object-centric models use slots and attention-based routing for binding. However, this class of models has several conceptual limitations: the number of slots is hardwired; all slots have equal capacity; training has high computational cost; there are no object-level relational factors within slots. Synchrony-based models in principle can address these limitations by using complex-valued activations which store binding information in their phase components. However, working examples of such synchrony-based models have been developed only very recently, and are still limited to toy grayscale datasets and simultaneous storage of less than three objects in practice. Here we introduce architectural modifications and a novel contrastive learning method that greatly improve the state-of-the-art synchrony-based model. For the first time, we obtain a class of synchrony-based models capable of discovering objects in an unsupervised manner in multi-object color datasets and simultaneously representing more than three objects.

PaperPDFConference PDFCodeCode Syntology ranMethod as code

In Syntology View this paper on Syntology, its page in Syntology's graph. That page adds which of its functions are also present in other papers' code Syntology holds, with dates, and its dated run record as JSON and a README badge. It also lists the repositories linked to the paper, the abstract and the calls for agents.

Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Syntology holds a reference implementation of Contrastive Training of Complex-Valued Autoencoders, generated from the paper text (not the authors' code), at level 3 on Syntology's ladder: V2 “two implementations GENERATED from the paper's signature and spec agree numerically (1e-6) on fixture inputs” “neither of the two ever read the authors' own code, so V2 measures how unambiguously the paper specifies the method, not fidelity to the authors' implementation (measured 2026-09-11, commit 75eaa77: cross_checked_by is a second LLM on 2,628 of 2,831 served samples)”; V3 “paper-derived property tests pass on both of those same two generations.” (quoted from Syntology's server card). No rung compares this code against the authors' implementation. For agents: get_reference_implementation(name='Contrastive Training of Complex-Valued Autoencoders') (how to connect); the free check have('2305.15001') needs no token.

For agents, on Syntology's MCP service (how to connect):

Code

Syntology 38 code samples harvested from 1 repository linked to this paper: 29 ran (of which 5 constructed an object rather than computing a result; 11 with no instrument failure: 4 honoured, 1 violated, 6 with no contract checked; 18 where Syntology's instrument failed) · 9 unverified. Where Syntology's instrument failed: 14 where our contract draft was wrong; 4 where the contract check could not run; failures of Syntology's instrument, not of the code.

By repository: official repository: 38 samples from 1 repository, 29 ran (11 with no instrument failure; 18 where Syntology's instrument failed). The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

Levels written between 2026-09-22 and 2026-09-24 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed. 3 of the 29 samples that ran have no recorded date.

agopal42/ctcae officialmentioned in papermentioned on GitHubpytorchSyntology: 38 samples, 29 ran (18 instrument-failed) · 9 unverified report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

38 samples harvested: 29 ran (of which 5 constructed an object rather than computing a result; 11 with no instrument failure: 4 honoured, 1 violated, 6 with no contract checked; 18 where Syntology's instrument failed) · 9 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page last changed ; when this build read Syntology's graph is in the build record. That date is not when the samples ran.

4ran · honoured contract
1ran · violated contract
14ran · our draft was wrong
4ran; the contract check could not run
6ran
9unverified

Licence: 0 of the 38 samples are pointer only, meaning Syntology does not serve that copy's text. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from agopal42/ctcae. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. A line beginning “fixture:” under a sample that ran gives the arguments Syntology's fixture called the sample with, each as name:type/rank/element kind (? where none was recorded), and the type of what the fixture's run returned, exactly as the harness recorded them; the arguments and the returned type were recorded by two separate runs of the same fixture. No output value is shown, and a returned type says nothing about whether the output was right. “Level recorded” gives the UTC date when Syntology's graph wrote the sample's verification level, not when the code executed; where the level was written before Syntology's graph kept that date, the row says “level recorded, date not recorded”. For 1 sample that ran, the graph records a signature this page does not show: for a class the fixture built without calling, the harness called the object with inputs it made up itself, so the recorded signature is the harness's probe, not the fixture's arguments. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself. “Our draft was wrong” and “the contract check could not run” are failures of Syntology's instrument, not of the code: the sample ran on Syntology's synthesized input, and then the drafted contract did not describe it, or the check itself raised an error (RAISES), lacked a dependency (DEP_MISSING) or ran out of time (TIMEOUT); hover the label for which. A metamorphic tier, where shown, is evidence of the output's shape, not that the sample implements the method.

Each sample ends with its code_sha256. A sample's code_sha256 is Syntology's identity for that exact code. It is printed in full (16 hexadecimal characters for every sample in this build), and get_code(code_sha256="…") takes it exactly as printed. An agent calls it on Syntology's MCP service (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted. community: Not in the archive's code links for this paper; a community repository Syntology harvested. Samples are grouped by repository: a repository marked official first, then by how directly the repository is tied to the paper, then in the archive's order; never by how many ran. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of six reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar); this paper's copy was not located, so its licence is not recorded; or this paper's copy is cleared, but the stored copy of the same code that Syntology serves is not. Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

agopal42/ctcae official repository 38 samples, 29 ran (18 instrument-failed) · 9 unverified
ComplexConv2d models/ctcae_model.py ran MIT (permissive) · b94524beed8991fd · report fixture: returned ComplexConv2d · level recorded 2026-09-22
ComplexConvTranspose2d models/ctcae_model.py ran MIT (permissive) · 87f9df0c5ba0263b · report fixture: returned ComplexConvTranspose2d · level recorded 2026-09-22
ComplexConvUpsample2d models/ctcae_model.py ran MIT (permissive) · 32fbb448bc6bee93 · report fixture: signature x:Tensor/4/complex · returned tuple · level recorded 2026-09-24
ComplexLinear models/ctcae_model.py ran MIT (permissive) · 89bd4fbb5bfe4b2e · report fixture: returned ComplexLinear · level recorded 2026-09-22
ComplexUpsample models/ctcae_model.py ran MIT (permissive) · 2d311e2e21a3dbbf · report fixture: returned ComplexUpsample · level recorded 2026-09-22
Conv2dTF models/ctcae_model.py ran fingerprinted MIT (permissive) · e7d7eb775a61658c · report fixture: returned Conv2dTF · level recorded 2026-09-22
append_black_right_separator models/ctcae_model.py ran; the contract check could not run (RAISES) fingerprinted MIT (permissive) · e38855884ac1991b · report fixture: signature imgs:ndarray/4/int · returned ndarray · level recorded 2026-09-22
append_border_pixels models/ctcae_model.py ran; the contract check could not run (RAISES) MIT (permissive) · d3123efe45c58cc9 · report fixture: signature imgs:ndarray/4/int, color:str/0/str · returned ndarray · level recorded 2026-09-22
apply_kmeans models/ctcae_model.py ran · our draft was wrong MIT (permissive) · 8f2a3483310f84a5 · report fixture: signature phase_mask_threshold:float/0/float, output_phase:Tensor/4/float, output_norm_magnitude:Tensor/3/float, labels_true:Tensor/3/int, img_resolution:tuple/1/int, seed:int/0/int · returned tuple · level recorded 2026-09-22
apply_layer models/ctcae_model.py ran · our draft was wrong MIT (permissive) · 193817cd55076b39 · report fixture: signature x:Tensor/2/complex, real_function:callable/0/?, phase_bias:Tensor/2/float, magnitude_bias:Tensor/2/float · returned tuple · level recorded 2026-09-22
batch_to_rowwise_image models/ctcae_model.py ran · violated contract MIT (permissive) · bfcf0002f0908e3b · report fixture: signature imgs:ndarray/4/float · returned ndarray · level recorded 2026-09-22
calc_ari_score models/ctcae_model.py ran · our draft was wrong fingerprinted MIT (permissive) · 8fbf469567396916 · report fixture: signature batch_size:int/0/int, labels_true:ndarray/3/int, labels_pred:ndarray/3/int, with_background:bool/0/bool · returned tuple · level recorded 2026-09-22
calculate_cluster_metrics models/ctcae_model.py ran · our draft was wrong MIT (permissive) · 82c5b64172909c18 · report fixture: signature in_phase:ndarray/2/float, k_means:KMeans/0/? · returned dict · level recorded 2026-09-22
clip_and_rescale models/ctcae_model.py ran · honoured contract fingerprinted MIT (permissive) · e30f758e4571b29b · report fixture: signature input_tensor:Tensor/2/float, clip_value:float/0/float · returned Tensor · level recorded 2026-09-22
color_hex_to_float models/ctcae_model.py ran · honoured contract MIT (permissive) · 6d8ef248dfd50c0e · report fixture: signature hex_color:str/0/str · returned ndarray · level recorded 2026-09-22
color_hex_to_int models/ctcae_model.py ran · honoured contract MIT (permissive) · 0f73b6f5607dbfc0 · report fixture: signature hex_color:str/0/str · returned ndarray · level recorded 2026-09-22
combined_labels_to_colormap models/ctcae_model.py ran · our draft was wrong fingerprinted MIT (permissive) · b1411375ee85340e · report fixture: signature combined_labels:Tensor/3/int, number_of_classes:int/0/int · returned ndarray · level recorded 2026-09-22
get_anchors_positive_and_negative_pairs models/ctcae_model.py ran; the contract check could not run (RAISES) fingerprinted MIT (permissive) · d8c2b4c45997c4de · report fixture: signature addresses:Tensor/4/float, features:Tensor/4/float, n_anchors_to_sample:int/0/int, n_positive_pairs:int/0/int, n_negative_pairs:int/0/int, top_k:int/0/int, bottom_m:int/0/int, use_patches:bool/0/bool, patch_size:int/0/int · returned tuple · level recorded 2026-09-22
get_ari_for_n_objects models/ctcae_model.py ran · our draft was wrong fingerprinted MIT (permissive) · 94ad841b3183e122 · report fixture: signature ari_w_bg_scores:ndarray/1/float, ari_wo_bg_scores:ndarray/1/float, n_objects:ndarray/1/int, n_objects_max:int/0/int · returned tuple · level recorded 2026-09-22
get_conv_biases models/ctcae_model.py ran · our draft was wrong fingerprinted MIT (permissive) · 48ac46b5f4b9ad19 · report fixture: signature n_out_channels:int/0/int, fan_in:int/0/int · returned tuple · level recorded, date not recorded
get_conv_layers_shapes models/ctcae_model.py ran · honoured contract MIT (permissive) · 336163d2a29c0496 · report fixture: signature conv_layers:callable/0/?, d_channels:int/0/int, d_height:int/0/int, d_width:int/0/int · returned list · level recorded 2026-09-22
get_stable_angle models/ctcae_model.py ran · our draft was wrong MIT (permissive) · 09227b2bfcaed35e · report fixture: signature x:Tensor/2/complex, eps:float/0/float · returned Tensor · level recorded 2026-09-22
init_conv_norms models/ctcae_model.py ran · our draft was wrong MIT (permissive) · d910d2285daa4de1 · report fixture: signature conv_shapes:list/1/tuple · returned ModuleList · level recorded, date not recorded
init_magnitude_bias models/ctcae_model.py ran · our draft was wrong fingerprinted MIT (permissive) · d97fbc4d2b3f3597 · report fixture: signature fan_in:int/0/int, bias:Tensor/1/float · returned Tensor · level recorded, date not recorded
init_phase_bias models/ctcae_model.py ran · our draft was wrong fingerprinted MIT (permissive) · 201529a35cce67ab · report fixture: signature bias:Tensor/1/float · returned Parameter · level recorded 2026-09-22
phase_to_cartesian_coordinates models/ctcae_model.py ran · our draft was wrong fingerprinted MIT (permissive) · f53d25628716495f · report fixture: signature phase_mask_threshold:float/0/float, phase:Tensor/2/float, norm_magnitude:Tensor/1/float · returned Tensor · level recorded 2026-09-22
resize_image_batch models/ctcae_model.py ran; the contract check could not run (RAISES) MIT (permissive) · 456f8ca4ca4808a4 · report fixture: signature image_batch:ndarray/4/int, new_size:tuple/1/int · returned ndarray · level recorded 2026-09-22
spherical_to_cartesian_coordinates models/ctcae_model.py ran · our draft was wrong fingerprinted MIT (permissive) · f10f65f89cbbf0d1 · report fixture: signature x:Tensor/2/float · returned Tensor · level recorded 2026-09-22
tensor_to_wandb_image models/ctcae_model.py ran · our draft was wrong MIT (permissive) · 419682da2a94001b · report fixture: returned Image · level recorded 2026-09-22
CAEMetricsRecorder models/ctcae_model.py unverified MIT (permissive) · c4b48ec9eac36ca1 · report
ComplexAutoEncoderV2 models/ctcae_model.py unverified MIT (permissive) · 9e50df740846fc6b · report
ComplexConvDecoder models/ctcae_model.py unverified MIT (permissive) · a9b5cc6c540885a0 · report
ComplexEncoder models/ctcae_model.py unverified MIT (permissive) · a1c42747d2fcef56 · report
append_mutiple_radial_plot_tuples models/ctcae_model.py unverified MIT (permissive) · bf4c0805a187d0c9 · report
append_single_radial_plot_tuple models/ctcae_model.py unverified MIT (permissive) · 54c7e6a22859d9a0 · report
canvas_to_numpy models/ctcae_model.py unverified MIT (permissive) · 42afe1af390dce9a · report
create_image_grids_for_logging models/ctcae_model.py unverified MIT (permissive) · f2983ab9dcaa0a6d · report
create_phase_colorcoded_groupings_and_radial_plots models/ctcae_model.py unverified MIT (permissive) · 3e77d62b6f031b43 · report

Method as code Syntology

What Syntology's graph records as this paper's claimed method, lifted into code: each assembly names its bricks (functions and classes), each with its own role and its own verification level. An assembly is not a check that the code implements the method correctly or reproduces the paper's results. Levels are each brick's own verification level in Syntology's graph (level 0: no recorded run), listed brick by brick and never combined into one level; “ran” means the brick executed on a synthesized input, not that its output is correct, and the instrument figure counts failures of Syntology's instrument, not of the code. For a generated assembly, an independent peer is a second implementation generated separately from the same paper text; neither implementation read the authors' code. At level 2 the two agreed numerically on the fixture's inputs; level 3 is recorded when property tests derived from the paper also passed on both, and each level 3 card prints how many its report lists. Where a model read the code against the paper's claim, each assembly's label verdict is one model's judgement, kept in this page's JSON twin; it is not calibrated, so no card prints it. A generated assembly's label was set when its code was generated for the method, and no model judged it; its card prints it as "by construction" or "not judged by a model". Where an assembly's phrase has no constructed figure, a brick that ran is not among the harvested samples this build reads, and whether a brick's run only constructed an object is read from that brick's harvested sample, so the figure was not counted; it does not say there were none.

2 assemblies on this paper: 1 harvested, 1 generated.

generated Assembly for Contrastive Training of Complex-Valued Autoencoders (graph method: Contrastive Training of Complex-Valued Autoencoders), generated from the paper text, not the repository's code. 1 brick; individually on a fixture: 1 ran (1 with no instrument failure: 0 honoured, 1 violated, 0 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified. Ran end-to-end against an independent peer (level 3; 5 property tests passed on both).

label matches by construction: the code was generated for this method (not judged by a model)

  1. cosine_distance entry · level 3 · ran · violated contract · c469701270d20ee7ce057bd72b9f2107cef145fabaf5c6f34e150034a099068d

harvested Assembly for Contrastive Training of Complex-Valued Autoencoders (graph method: Contrastive Training of Complex-Valued Autoencoders). 38 bricks; individually on a fixture: 29 ran (of which 5 constructed an object rather than computing a result; 11 with no instrument failure: 4 honoured, 1 violated, 6 with no contract checked; 18 where Syntology's instrument failed) · 9 unverified. The assembly has not been run end-to-end.

  1. ComplexAutoEncoderV2 entry · level 0 · unverified · 9e50df740846fc6b
  2. CAEMetricsRecorder helper · level 0 · unverified · c4b48ec9eac36ca1
  3. ComplexConvDecoder model · level 0 · unverified · a9b5cc6c540885a0
  4. ComplexEncoder model · level 0 · unverified · a1c42747d2fcef56
  5. get_anchors_positive_and_negative_pairs helper · level 1 · ran; the contract check could not run (RAISES) · d8c2b4c45997c4de
  6. apply_kmeans helper · level 1 · ran · our draft was wrong · 8f2a3483310f84a5
  7. calc_ari_score helper · level 1 · ran · our draft was wrong · 8fbf469567396916
  8. clip_and_rescale helper · level 1 · ran · honoured contract · e30f758e4571b29b
  9. create_image_grids_for_logging helper · level 0 · unverified · f2983ab9dcaa0a6d
  10. create_phase_colorcoded_groupings_and_radial_plots helper · level 0 · unverified · 3e77d62b6f031b43
  11. get_ari_for_n_objects helper · level 1 · ran · our draft was wrong · 94ad841b3183e122
  12. tensor_to_wandb_image helper · level 1 · ran · our draft was wrong · 419682da2a94001b
  13. ComplexConv2d model · level 1 · ran · b94524beed8991fd
  14. ComplexConvTranspose2d model · level 1 · ran · 87f9df0c5ba0263b
  15. ComplexConvUpsample2d model · level 1 · ran · 32fbb448bc6bee93
  16. ComplexLinear model · level 1 · ran · 89bd4fbb5bfe4b2e
  17. get_conv_layers_shapes helper · level 1 · ran · honoured contract · 336163d2a29c0496
  18. init_conv_norms helper · level 1 · ran · our draft was wrong · d910d2285daa4de1
  19. calculate_cluster_metrics helper · level 1 · ran · our draft was wrong · 82c5b64172909c18
  20. phase_to_cartesian_coordinates helper · level 1 · ran · our draft was wrong · f53d25628716495f
  21. append_black_right_separator helper · level 1 · ran; the contract check could not run (RAISES) · e38855884ac1991b
  22. append_border_pixels helper · level 1 · ran; the contract check could not run (RAISES) · d3123efe45c58cc9
  23. append_mutiple_radial_plot_tuples helper · level 0 · unverified · bf4c0805a187d0c9
  24. append_single_radial_plot_tuple helper · level 0 · unverified · 54c7e6a22859d9a0
  25. batch_to_rowwise_image helper · level 1 · ran · violated contract · bfcf0002f0908e3b
  26. combined_labels_to_colormap helper · level 1 · ran · our draft was wrong · b1411375ee85340e
  27. resize_image_batch helper · level 1 · ran; the contract check could not run (RAISES) · 456f8ca4ca4808a4
  28. canvas_to_numpy helper · level 0 · unverified · 42afe1af390dce9a
  29. Conv2dTF helper · level 1 · ran · e7d7eb775a61658c
  30. apply_layer model · level 1 · ran · our draft was wrong · 193817cd55076b39
  31. get_conv_biases helper · level 1 · ran · our draft was wrong · 48ac46b5f4b9ad19
  32. ComplexUpsample model · level 1 · ran · 2d311e2e21a3dbbf
  33. init_magnitude_bias helper · level 1 · ran · our draft was wrong · d97fbc4d2b3f3597
  34. init_phase_bias helper · level 1 · ran · our draft was wrong · 201529a35cce67ab
  35. spherical_to_cartesian_coordinates helper · level 1 · ran · our draft was wrong · f10f65f89cbbf0d1
  36. color_hex_to_float helper · level 1 · ran · honoured contract · 6d8ef248dfd50c0e
  37. get_stable_angle helper · level 1 · ran · our draft was wrong · 09227b2bfcaed35e
  38. color_hex_to_int helper · level 1 · ran · honoured contract · 0f73b6f5607dbfc0

Each brick ends with its code_sha256, Syntology's identity for that exact code; click one to copy the get_code call that returns it. For agents, on Syntology's MCP service: assembly(arxiv_id="2305.15001") returns these assemblies with each brick's role and wiring (how to connect).

Tasks

Contrastive LearningObjectObject Discovery

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Methods

Contrastive Learning

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections