SEMA: Exploring a Different AI Architecture

I believe AI should remain transparent, inspectable, and accessible to independent researchers—not only to organizations with massive compute budgets. SEMA is motivated by the idea that architectural diversity matters: AI should support local execution, data sovereignty, explicit reasoning, and source availability, ensuring that research remains reproducible, understandable, and independent of centralized infrastructure.

I also believe that AI should not become synonymous with a single architectural paradigm. The extraordinary success of LLMs should inspire further exploration, not diminish investment in alternative approaches. A healthy scientific ecosystem depends on architectural diversity, where fundamentally different ideas continue to be explored, challenged, and improved alongside the dominant paradigm.

I’ve been working on SEMA, an experimental AI architecture that explores a direction fundamentally different from transformer-based LLMs.

The goal is not to build a smaller LLM.

The goal is to investigate whether language understanding and reasoning can emerge from explicit semantic structure instead of next-token prediction.

Some characteristics of SEMA:

Not transformer-based.

Not an LLM.

CPU-first design.

Multilingual by construction.

Transparent internal representations instead of opaque embeddings.

Explicit relational memory rather than storing knowledge only in neural weights.

Composable semantic representations using Vector Symbolic Architecture (VSA/HDC).

Graph-based reasoning instead of autoregressive generation.

One design decision I find particularly interesting is separating semantic representation from knowledge storage.

Instead of trying to encode everything into dense neural parameters, SEMA stores explicit relational knowledge and uses high-dimensional semantic representations as an indexing and reasoning mechanism over that knowledge.

This makes it possible to inspect structures, analyze intermediate representations, and experiment with reasoning algorithms directly.

The project is source available because I want researchers to study, reproduce, and challenge the architecture while protecting the continuity of independent research.

I’m not claiming this approach is superior to LLMs.

I’m asking a different question:

Are there still unexplored architectural directions that deserve serious research alongside the transformer ecosystem?

I’d greatly appreciate feedback from anyone interested in alternative AI architectures, VSA/HDC, cognitive systems, graph reasoning, or symbolic approaches.

Repository: GitHub - hviana/sema: Sema is a non-parametric, instance-based reasoning system: a Vector Symbolic Architecture over a content-addressable memory, with inference by weighted automated deduction. · GitHub

Hello,

I am not an expert, but I am very interested in contributing and learning from the project.

I read the repository and your explanation of SEMA. As I understand it, SEMA is a non-parametric, instance-based reasoning architecture that combines Vector Symbolic Architecture / Hyperdimensional Computing with content-addressable relational memory and weighted automated deduction.

What I find especially interesting is that it does not treat knowledge as something hidden inside billions of neural parameters. Instead, it keeps knowledge explicit, inspectable, and traceable, while using vector-symbolic representations to retrieve related structures and weighted deduction to compose derivations.

I see SEMA as more than only a pattern matcher: it appears to combine approximate structural retrieval, analogy-like generalization, explicit memory, and multi-step graph-based reasoning. The deterministic and auditable design is especially valuable for tasks where it is important to know why a system produced an answer.

My question is about scalability and open-ended reasoning. Symbolic systems can perform planning and multi-step reasoning, but in broad real-world settings they may face challenges such as search-space explosion, ambiguity in natural language, noisy or incomplete knowledge, and the difficulty of extracting reliable structured facts from ordinary text.

Because of that, I think SEMA could be especially powerful as part of a hybrid architecture. For example, a neural model such as a Transformer could handle language, perception, abstraction, and possibly high-level planning, while SEMA could provide explicit relational memory, deterministic computation, traceable deduction, and verifiable reasoning.

This could create a useful separation between knowledge and reasoning: learned neural components could interpret unstructured inputs, while the symbolic memory system could store explicit facts and support inspectable derivations over them.

Best regards,
Bidram

Regarding your question about the Transformer ecosystem, my personal view is that the field has built an enormous ecosystem around an architecture with important limitations. I am not saying that Transformers are bad, they have achieved remarkable results, but they are also expensive, computationally intensive, and often inefficient for tasks involving long-term memory, continual updating, or fully traceable reasoning. It’s like trying to build a full, native-speed operating system entirely in Python.

I sometimes wonder whether the strong focus on scaling Transformers has reduced the attention given to potentially better or more efficient architectural directions. Exploring alternatives such as symbolic, vector-symbolic, memory-based, or hybrid neural-symbolic systems could lead to architectures with different and valuable trade-offs.

For now, I tried partially testing it in Colab:


I think the direct answer to the broader question is yes: architectural directions outside the Transformer/LLM ecosystem still deserve serious investigation.

SEMA is particularly interesting because it does not merely replace one neural component with another. It combines explicit content-addressed memory, vector-symbolic similarity, instance-based transfer, weighted deduction, exact computation, and inspectable traces in one system.

Since the thread already contains a useful discussion of possible hybrid architectures, I focused on a narrower question:

What would help distinguish the capabilities already demonstrated by SEMA from the boundaries that still need to be measured?

I ran two small diagnostic passes against the pinned v0.2.4 release, at commit d14e55c, using Node 22.16.0 on a CPU-only Colab runtime.

The documented demo reproduced successfully, as did the selected upstream regression tests I ran. The tests below are not a general benchmark, and they are not intended as a verdict on the architecture. They only probe the local boundary around the current painter example.

My main result was:

The example clearly demonstrates structural answer-slot transfer followed by graph continuation. In this fixture, however, predicate-sensitive relation abstraction was not yet separated from those two capabilities.

That distinction seems useful because the documented result combines several operations in one successful answer:

  1. recognize a structurally similar stored case;
  2. identify an answer-like span;
  3. ground that span to Pablo Picasso;
  4. follow the explicit memory edge from Picasso to another fact;
  5. compose and articulate the final answer.

The full pipeline works on the positive example. The interesting next step is to determine exactly which of these operations depends on the learned relation, which depends on positional or structural similarity, and which depends on later graph traversal.

What reproduced

With the documented deposits:

The Mona Lisa was painted by Leonardo da Vinci. → Leonardo da Vinci
The Starry Night was painted by Vincent van Gogh. → Vincent van Gogh
The Night Watch was painted by Rembrandt van Rijn. → Rembrandt van Rijn
Pablo Picasso → Pablo Picasso co-founded the Cubist movement

the unseen query:

The Weeping Woman was painted by Pablo Picasso.

produced:

Pablo Picasso co-founded the Cubist movement

This matches the documented demonstration.

The strongest positive result from the diagnostic pass was not only that the answer appeared, but that the two-stage dependency behaved causally when the stored evidence was changed:

Intervention Observed result
Painter examples + Picasso continuation present Full Picasso continuation
Picasso continuation removed Pablo Picasso only
Painter examples removed The sentence query became silent
Painter examples removed, but Pablo Picasso queried directly Atomic continuation still worked
Picasso continuation replaced The answer followed the replacement

This gives reasonably direct evidence for the following path in this fixture:

sentence
  → entity/slot grounding
  → Pablo Picasso
  → explicit continuation edge
  → final fact

That is a useful result. It shows more than memorizing the complete final sentence, and it is compatible with the architecture described in HOW_IT_WORKS.md.

Where a small contrast set helped

I then changed only the relation wording while preserving the same sentence shape and person:

The Weeping Woman was restored by Pablo Picasso.
The Weeping Woman was inspired by Pablo Picasso.
The Weeping Woman was photographed by Pablo Picasso.
The Weeping Woman was painted near Pablo Picasso.
The Weeping Woman was not painted by Pablo Picasso.

In this fixture, these variants still grounded Picasso and usually continued to the stored Cubist fact.

When I removed the Picasso continuation and tested only the slot extraction behavior, the following all returned Pablo Picasso:

The Weeping Woman was painted by Pablo Picasso.
The Weeping Woman was restored by Pablo Picasso.
The Weeping Woman was not painted by Pablo Picasso.

This does not establish that SEMA cannot represent or learn predicate distinctions under another curriculum, representation, or configuration.

It does suggest that the current headline example, by itself, does not distinguish among several possible explanations:

  • learning the relation painted by;
  • learning a broader verb + by + person frame;
  • identifying the differing suffix/answer span;
  • transferring a structurally aligned slot from similar examples;
  • applying one of several redundant grounding mechanisms that happen to agree on the positive example.

A few carefully selected contrasts may therefore be more informative than adding more same-template positive examples.

A practical minimal set could be:

X was painted by Y.
X was restored by Y.
X was not painted by Y.
Y painted X.
X was restored by A and painted by B.
Who painted X?

These separate at least four capabilities:

  • structural slot transfer;
  • predicate sensitivity;
  • argument-role sensitivity;
  • question/declarative transformation.

This is close to the motivation behind Contrast Sets and CheckList: rather than relying only on aggregate success, make small meaning-preserving or meaning-changing edits and inspect the local decision boundary.

Suggested default route

The lowest-cost, highest-information route seems to be:

  1. Pin one release, seed, fixture, deposit order, and runtime.
  2. Keep the existing positive example.
  3. Add 10–20 local contrast cases.
  4. Record the answer, provenance, and compact rationale for each case.
  5. Separate grounding, atomic continuation, and full composition.
  6. Delete or replace the stated supporting deposit and replay the query.
  7. Only then use narrow mechanism ablations where attribution is still ambiguous.

A public fixture could be very small. JSONL or CSV would be enough:

id
capability
deposits
query
expected_class
expected_relation_or_role
actual_answer
provenance
compact_rationale
version
commit
runtime
seed
deposit_order
store_or_fixture_checksum

Useful expected_class values might be:

exact_answer
acceptable_answer
justified_abstention
unsupported_answer

It may also help to score the relation or role separately from the final text. A system can sometimes reach the correct entity through the wrong relation, and final-answer matching alone would hide that distinction.

The repository already contains detailed regression fixtures, including a particularly useful test where a weakly grounded comparison previously returned an unrelated haiku continuation and the corrected behavior is “honest silence.” That test is a good precedent for this style of boundary case:

test/50-cast-analog-consensus-floor.test.mjs

So this may not require creating a new evaluation culture. It may mainly require organizing some existing and new regression cases by capability and expected outcome.

Detailed observations from the local contrast probes

1. Number of worked examples

I tested the positive painter query with one, two, and three painter examples.

Number of painter examples Observed output
1 Picasso continuation
2 Leonardo fragment + Picasso continuation
3 Picasso continuation

In this small fixture, the headline answer was therefore already obtainable from one worked example.

That does not show that three examples are generally unnecessary. It means only that the final answer alone cannot establish that this particular output depended on induction over all three examples.

The two-example condition was also informative because it produced a fused answer rather than a monotonic improvement. That suggests example count and the set of competing attention points may deserve their own stability test.

2. Active voice and argument roles

The passive input:

The Weeping Woman was painted by Pablo Picasso.

worked as documented.

The active-voice version:

Pablo Picasso painted The Weeping Woman.

produced:

Pablo Picasso painted by Leonardo da Vinci.

The trace indicated that this involved a different grounding/substitution path and a connector miss.

I would treat this as a useful argument-role boundary case, not as a broad conclusion about active voice. A compact role test could hold the entities and relation constant while varying only:

X was painted by Y.
Y painted X.
A told B that Y painted X.
X was restored by A and painted by Y.

3. Question form

The declarative query:

The Weeping Woman was painted by Pablo Picasso.

returned the Picasso continuation.

The question:

Who painted The Weeping Woman?

was silent in this fixture.

That may be entirely consistent with the supplied training format, which teaches context-to-continuation mappings rather than an explicit declarative-to-interrogative transformation.

It may therefore be useful to treat these as different capability levels:

declarative completion
relation-sensitive extraction
question transformation
open-ended question answering

4. Unknown entity

For:

The Weeping Woman was painted by Alice Example.

the system returned:

Alice Example

rather than silence.

This appears consistent with successful answer-slot extraction followed by the absence of a known continuation.

For evaluation purposes, it would be useful to distinguish:

  • correct slot extraction;
  • successful stored-fact continuation;
  • complete task answer;
  • justified abstention.

Otherwise, a partial intermediate result may be counted either as success or failure depending on the evaluator’s unstated expectation.

5. Negation and fusion

For:

The Weeping Woman was not painted by Pablo Picasso.

one standard condition returned:

Leonardo da VinciPablo Picasso co-founded the Cubist movement

The trace separated this into approximately:

CAST grounds the Picasso continuation
    +
another attention point contributes a Leonardo fragment
    +
fusion combines them
    +
no connector is found
    +
bare concatenation is emitted

I then tested a diagnostic build with only the downstream fusion step disabled.

The extra Leonardo fragment disappeared, but the positive Picasso continuation remained.

That suggests two separable boundaries:

  1. an upstream relation/negation grounding issue;
  2. a downstream fusion/articulation issue that can amplify the malformed result.

This is a good example of why the trace is valuable. It permits a malformed answer to be localized instead of treating the whole system as one opaque pass.

The fusion-disabled condition was only a diagnostic patch to a copied compiled build, not a supported configuration or proposed fix.

6. Redundant mechanisms

The headline positive answer could be produced through more than one route.

  • With CAST disabled, extraction could ground Picasso and reasoning could continue.
  • With extraction disabled, CAST could produce the route.
  • With both removed, recall could still supply a related route in some conditions.
  • Removing an individual mechanism therefore did not necessarily remove the headline capability.

This may be desirable redundancy, but it complicates attribution.

Ablation conclusions should therefore be phrased carefully:

“The output survived removal of mechanism A”

does not necessarily mean:

“Mechanism A contributes nothing.”

It may mean that another mechanism implements an overlapping capability.

Combined ablations are useful only after the behavioral boundary is clear; otherwise they can become a search through implementation combinations without a stable capability definition.

7. Explicit relation-contrast deposits

I also tried a small exploratory condition where painted, restored, and not painted examples were deposited with different intended continuations.

In that particular setup, the tested relation-specific expectations were not recovered.

This should not be read as proof that relation-specific learning is impossible in SEMA.

It only shows that ordinary context-to-continuation deposits in that small curriculum did not immediately disambiguate the relations. Other possibilities remain open:

  • a different example curriculum;
  • explicit relation nodes;
  • a representation designed for predicate/argument structure;
  • additional negative or contrastive evidence;
  • a different mechanism boundary;
  • a future dedicated interface for constraints or negative cases.

The relevant comparison concept here may be anti-unification, which studies how to extract a least-general common structure from examples.

That is not a claim that SEMA is “just anti-unification.” It is a useful baseline question:

What part of the demonstrated transfer requires SEMA’s vector-symbolic geometry, and what part could also be explained by ordinary common-structure extraction?

A capability ladder that may help organize future tests

The broad phrase “language understanding and reasoning” contains several separable capabilities.

A possible ladder is:

Level Capability Minimal diagnostic
0 Exact memory / atomic retrieval Query a stored key directly
1 Structural answer-slot transfer New entity in a familiar frame
2 Predicate sensitivity painted versus restored
3 Argument-role sensitivity Passive versus active; two entities
4 Negation, unknown, and conflict Positive, negative, unsupported, contradictory
5 Multi-hop composition Ground an intermediate and continue
6 Memory lifecycle Corpus growth, correction, retraction, stale knowledge

The current painter fixture gives useful evidence for levels 0 and 1, and for part of level 5.

The local contrast probes suggest that levels 2–4 should be measured independently rather than inferred from the positive level-5 output.

This also prevents a system from being judged too harshly or too generously:

  • failing question transformation does not erase successful structural transfer;
  • successful entity extraction does not by itself establish predicate understanding;
  • successful atomic retrieval does not establish composition;
  • successful composition in one template does not establish role or negation handling.
Suggested decision tree
Can the pinned documented example be reproduced?
|
+-- No
|   |
|   +-- First fix the reproduction tuple:
|       version, commit, runtime, seed, fixture,
|       deposit order, configuration, expected trace
|
+-- Yes
    |
    +-- Run predicate, negation, and role contrasts
        |
        +-- Wrong entity or span
        |   |
        |   +-- Inspect recognition, CAST, extraction,
        |       halo evidence, and attention roots
        |
        +-- Correct entity, wrong continuation
        |   |
        |   +-- Inspect edge selection, recall,
        |       conflict resolution, and stored continuations
        |
        +-- Atomic stages succeed, combined answer fails
        |   |
        |   +-- Inspect multi-hop, fusion,
        |       connector selection, and articulation
        |
        +-- Result changes only with broader candidate search
        |   |
        |   +-- Investigate candidate-retrieval/index recall
        |
        +-- System is silent on nearly everything
        |   |
        |   +-- Measure answered-query risk together with coverage
        |
        +-- Local contrasts pass
            |
            +-- Move to held-out templates, longer hops,
                unrelated facts, contradictions,
                corpus growth, correction, and retraction

The default path would be to stop after the first few branches until each failure can be localized. A large benchmark becomes more useful after the expected behavior of these local cases is explicit.

How I would interpret provenance and auditability

SEMA’s trace system was one of the most useful parts of the experiment.

For future evaluation, I would separate four properties:

1. Execution consistency

Does the trace accurately record the mechanism and edges that actually ran?

2. Support sufficiency

Are the displayed deposits and derivation steps sufficient to support the answer?

3. Causal necessity

If a displayed supporting deposit is deleted or replaced, does the answer change in the predicted way?

4. Minimality

Does the rationale avoid presenting unrelated evidence as necessary support?

The evidence deletion and replacement probes gave encouraging results for causal sensitivity in the painter fixture.

This distinction may be helpful because “a trace exists” and “the trace is a faithful explanation” are not exactly the same claim.

A small replayable fixture is probably more informative here than a prose explanation alone:

original deposits
query
answer
trace

minus supporting deposit
same query
new answer

replacement supporting deposit
same query
replacement-following answer
Abstention and unsupported answers

The repository already treats “honest silence” as an important target, which seems appropriate for this architecture.

It may help to report several outcomes separately:

correct answer
justified abstention
excessive abstention
false grounding
retrieval omission
unsupported composition
near-match or exemplar echo

A low unsupported-answer rate can always be obtained by answering fewer queries, so abstention quality should be evaluated together with answer coverage.

The selective-prediction literature discusses this as a risk/coverage trade-off; a useful recent overview is On the Evaluation of Neural Selective Prediction Methods for Natural Language Processing.

The methodology is applicable even though SEMA is not a neural selective-prediction model:

  • vary a confidence or grounding threshold;
  • measure how many queries receive answers;
  • measure error among answered queries;
  • separately count cases where the system remained silent despite sufficient stored evidence.
Later-stage comparisons, after the local boundary is clearer

These do not all need to be done now. They are possible next branches after the small contrast suite becomes stable.

VSA contribution

VSA/HDC is a family of representations and binding operations, not one uniform algorithm. Capacity, bundling, and approximate unbinding behavior differ among variants.

A useful comparison would be:

current SEMA representation
independent random byte vectors
character n-gram retrieval
edit-distance nearest case
exact graph-only retrieval
simple common-structure extraction

The question would not be whether VSA is “good” or “bad,” but:

Which tested capability disappears or degrades when the VSA-specific component is replaced?

A broad comparison of VSA families is available in A Comparison of Vector Symbolic Architectures.

Longer relational composition

Once local role and predicate contrasts work reliably, benchmarks such as:

  • CLUTRR, for relation-chain generalization and distracting facts;
  • ProofWriter, for proof depth and True/False/Unknown distinctions;
  • COGS, for familiar components in unseen structural combinations;

could provide useful design ideas.

I would borrow their evaluation dimensions before attempting a full benchmark port.

Candidate retrieval and scaling

The exact graph can only verify candidates that the approximate index has surfaced.

At larger scale, useful measurements would therefore include:

candidate recall at k
final-answer agreement with a reference search
abstention agreement
trace agreement
query latency
index size
index build time

ANN-Benchmarks is useful for the general recall/speed methodology, although an end-to-end SEMA evaluation also needs to measure whether candidate differences change the final derivation.

Multilingual behavior

“Multilingual” may be easier to evaluate if divided into:

arbitrary UTF-8 storage
script independence
tokenizer-vocabulary independence
explicit cross-language aliases
parallel-pair transfer
shared non-text grounding
zero-link cross-language generalization

These are different capabilities.

Byte-addressed identity also raises a practical distinction between exact stored bytes and user-facing string equivalence. The W3C Character Model for the World Wide Web: String Matching and Searching gives useful background on Unicode normalization and equivalent-looking strings.

Memory lifecycle

A long-running explicit memory eventually needs policies for:

correction
retraction
contradictory sources
duplicate evidence
source independence
time-bounded facts
derived-fact invalidation

This is not specific to SEMA. It is a standard concern in case-based and explicit knowledge systems.

The classic retrieve/reuse/revise/retain cycle from case-based reasoning may provide useful vocabulary without replacing SEMA’s own terminology.

A compact proposed public probe

A very small first public suite could look like this:

Group Example Expected distinction
Positive control X was painted by Y Structural transfer
Predicate contrast X was restored by Y Different relation
Negation X was not painted by Y Do not preserve positive entailment
Role reversal Y painted X Preserve argument roles
Two relations X was restored by A and painted by B Select the correct role
Unknown Known frame, unknown continuation Partial extraction or abstention
Atomic hop Query Y directly Verify continuation
Composition Frame query leading to Y Verify two-stage path
Evidence deletion Remove Y → fact Full answer should disappear
Evidence replacement Replace Y → fact Answer should follow replacement
Irrelevant additions Add unrelated deposits Positive result should remain stable

For each test, recording a compact trace and expected outcome class would make the result useful to both human reviewers and automated regression tooling.

Overall impression

The experiment left me with a more positive and more specific view than either “it works” or “it fails.”

What seems demonstrated in the current painter fixture is:

  • explicit atomic memory;
  • structural case/slot transfer;
  • continuation through an explicit stored relation;
  • useful internal traces;
  • causal sensitivity to supporting-deposit deletion and replacement;
  • exact arithmetic through the separate ALU path.

What remains useful to isolate is:

  • predicate-sensitive abstraction;
  • argument-role preservation;
  • negation and unknown handling;
  • mechanism attribution when several routes are redundant;
  • stability under competing examples and corpus growth;
  • calibrated silence versus unsupported answers.

That seems like a productive position for an experimental architecture.

The components have substantial precedent individually, but the interesting research question is what capabilities emerge from this particular integration—and under what minimal conditions they stop emerging.

A small, versioned, trace-bearing contrast suite around the existing regression tests may be the most efficient next artifact for making that boundary visible to other researchers.

Thank you for the careful analysis. Your description captures SEMA well, although I would not frame it primarily as a symbolic component for a Transformer-based system. SEMA is an attempt to determine how far an architecture built from explicit relational memory, VSA-based structural representations, approximate analogy, and weighted graph derivation can go independently.

Scalability and combinatorial growth are central concerns, but SEMA does not approach reasoning as unrestricted symbolic expansion. It implements explicit controls over which structures may participate in a derivation and how competing paths are retained, combined, or discarded. These mechanisms are part of the architecture itself rather than an external search procedure.

A hybrid system is certainly possible, but the research question behind SEMA is more fundamental: whether language interpretation, generalization, memory, and multi-step reasoning can be supported by a transparent non-parametric architecture without making an LLM responsible for the difficult parts.