QuantumBoltzmannMachine

quantrs2_ml::boltzmann

Struct QuantumBoltzmannMachine 

Source 
pub struct QuantumBoltzmannMachine { /* private fields */ }
Expand description

Quantum Boltzmann Machine

Implementations§

Source§

impl QuantumBoltzmannMachine

Source

pub fn new( num_visible: usize, num_hidden: usize, temperature: f64, learning_rate: f64, ) -> Result<Self>

Create a new Quantum Boltzmann Machine

Examples found in repository?
examples/quantum_boltzmann.rs (lines 49-54)
47fn basic_qbm_demo() -> Result<()> {
48    // Create a small QBM
49    let mut qbm = QuantumBoltzmannMachine::new(
50        4,    // visible units
51        2,    // hidden units
52        1.0,  // temperature
53        0.01, // learning rate
54    )?;
55
56    println!("   Created QBM with 4 visible and 2 hidden units");
57
58    // Generate synthetic binary data
59    let data = generate_binary_patterns(100, 4);
60
61    // Train the QBM
62    println!("   Training on binary patterns...");
63    let losses = qbm.train(&data, 50, 10)?;
64
65    println!("   Training complete:");
66    println!("   - Initial loss: {:.4}", losses[0]);
67    println!("   - Final loss: {:.4}", losses.last().unwrap());
68
69    // Sample from trained model
70    let samples = qbm.sample(5)?;
71    println!("\n   Generated samples:");
72    for (i, sample) in samples.outer_iter().enumerate() {
73        print!("   Sample {}: [", i + 1);
74        for val in sample {
75            print!("{val:.0} ");
76        }
77        println!("]");
78    }
79
80    Ok(())
81}
82
83/// RBM demonstration with persistent contrastive divergence
84fn rbm_demo() -> Result<()> {
85    // Create RBM with annealing
86    let annealing = AnnealingSchedule::new(2.0, 0.5, 100);
87
88    let mut rbm = QuantumRBM::new(
89        6,    // visible units
90        3,    // hidden units
91        2.0,  // initial temperature
92        0.01, // learning rate
93    )?
94    .with_annealing(annealing);
95
96    println!("   Created Quantum RBM with annealing schedule");
97
98    // Generate correlated binary data
99    let data = generate_correlated_data(200, 6);
100
101    // Train with PCD
102    println!("   Training with Persistent Contrastive Divergence...");
103    let losses = rbm.train_pcd(
104        &data, 100, // epochs
105        20,  // batch size
106        50,  // persistent chains
107    )?;
108
109    // Analyze training
110    let improvement = (losses[0] - losses.last().unwrap()) / losses[0] * 100.0;
111    println!("   Training statistics:");
112    println!("   - Loss reduction: {improvement:.1}%");
113    println!("   - Final temperature: 0.5");
114
115    // Test reconstruction
116    let test_data = data.slice(s![0..5, ..]).to_owned();
117    let reconstructed = rbm.qbm().reconstruct(&test_data)?;
118
119    println!("\n   Reconstruction quality:");
120    for i in 0..3 {
121        print!("   Original:      [");
122        for val in test_data.row(i) {
123            print!("{val:.0} ");
124        }
125        print!("]  →  Reconstructed: [");
126        for val in reconstructed.row(i) {
127            print!("{val:.0} ");
128        }
129        println!("]");
130    }
131
132    Ok(())
133}
134
135/// Deep Boltzmann Machine demonstration
136fn deep_boltzmann_demo() -> Result<()> {
137    // Create a 3-layer DBM
138    let layer_sizes = vec![8, 4, 2];
139    let mut dbm = DeepBoltzmannMachine::new(
140        layer_sizes.clone(),
141        1.0,  // temperature
142        0.01, // learning rate
143    )?;
144
145    println!("   Created Deep Boltzmann Machine:");
146    println!("   - Architecture: {layer_sizes:?}");
147    println!("   - Total layers: {}", dbm.rbms().len());
148
149    // Generate hierarchical data
150    let data = generate_hierarchical_data(300, 8);
151
152    // Layer-wise pretraining
153    println!("\n   Performing layer-wise pretraining...");
154    dbm.pretrain(
155        &data, 50, // epochs per layer
156        30, // batch size
157    )?;
158
159    println!("\n   Pretraining complete!");
160    println!("   Each layer learned increasingly abstract features");
161
162    Ok(())
163}
164
165/// Energy landscape visualization
166fn energy_landscape_demo() -> Result<()> {
167    // Create small QBM for visualization
168    let qbm = QuantumBoltzmannMachine::new(
169        2,    // visible units (for 2D visualization)
170        1,    // hidden unit
171        0.5,  // temperature
172        0.01, // learning rate
173    )?;
174
175    println!("   Analyzing energy landscape of 2-unit system");
176
177    // Compute energy for all 4 possible states
178    let states = [
179        Array1::from_vec(vec![0.0, 0.0]),
180        Array1::from_vec(vec![0.0, 1.0]),
181        Array1::from_vec(vec![1.0, 0.0]),
182        Array1::from_vec(vec![1.0, 1.0]),
183    ];
184
185    println!("\n   State energies:");
186    for (i, state) in states.iter().enumerate() {
187        let energy = qbm.energy(state);
188        let prob = (-energy / qbm.temperature()).exp();
189        println!(
190            "   State [{:.0}, {:.0}]: E = {:.3}, P ∝ {:.3}",
191            state[0], state[1], energy, prob
192        );
193    }
194
195    // Show coupling matrix
196    println!("\n   Coupling matrix:");
197    for i in 0..3 {
198        print!("   [");
199        for j in 0..3 {
200            print!("{:6.3} ", qbm.couplings()[[i, j]]);
201        }
202        println!("]");
203    }
204
205    Ok(())
206}
Source

pub fn energy(&self, state: &Array1<f64>) -> f64

Compute energy of a configuration

Examples found in repository?
examples/quantum_boltzmann.rs (line 187)
166fn energy_landscape_demo() -> Result<()> {
167    // Create small QBM for visualization
168    let qbm = QuantumBoltzmannMachine::new(
169        2,    // visible units (for 2D visualization)
170        1,    // hidden unit
171        0.5,  // temperature
172        0.01, // learning rate
173    )?;
174
175    println!("   Analyzing energy landscape of 2-unit system");
176
177    // Compute energy for all 4 possible states
178    let states = [
179        Array1::from_vec(vec![0.0, 0.0]),
180        Array1::from_vec(vec![0.0, 1.0]),
181        Array1::from_vec(vec![1.0, 0.0]),
182        Array1::from_vec(vec![1.0, 1.0]),
183    ];
184
185    println!("\n   State energies:");
186    for (i, state) in states.iter().enumerate() {
187        let energy = qbm.energy(state);
188        let prob = (-energy / qbm.temperature()).exp();
189        println!(
190            "   State [{:.0}, {:.0}]: E = {:.3}, P ∝ {:.3}",
191            state[0], state[1], energy, prob
192        );
193    }
194
195    // Show coupling matrix
196    println!("\n   Coupling matrix:");
197    for i in 0..3 {
198        print!("   [");
199        for j in 0..3 {
200            print!("{:6.3} ", qbm.couplings()[[i, j]]);
201        }
202        println!("]");
203    }
204
205    Ok(())
206}
Source

pub fn create_gibbs_circuit(&self) -> Result<()>

Create quantum circuit for Gibbs state preparation

Source

pub fn sample(&self, num_samples: usize) -> Result<Array2<f64>>

Sample from the Boltzmann distribution

Examples found in repository?
examples/quantum_boltzmann.rs (line 70)
47fn basic_qbm_demo() -> Result<()> {
48    // Create a small QBM
49    let mut qbm = QuantumBoltzmannMachine::new(
50        4,    // visible units
51        2,    // hidden units
52        1.0,  // temperature
53        0.01, // learning rate
54    )?;
55
56    println!("   Created QBM with 4 visible and 2 hidden units");
57
58    // Generate synthetic binary data
59    let data = generate_binary_patterns(100, 4);
60
61    // Train the QBM
62    println!("   Training on binary patterns...");
63    let losses = qbm.train(&data, 50, 10)?;
64
65    println!("   Training complete:");
66    println!("   - Initial loss: {:.4}", losses[0]);
67    println!("   - Final loss: {:.4}", losses.last().unwrap());
68
69    // Sample from trained model
70    let samples = qbm.sample(5)?;
71    println!("\n   Generated samples:");
72    for (i, sample) in samples.outer_iter().enumerate() {
73        print!("   Sample {}: [", i + 1);
74        for val in sample {
75            print!("{val:.0} ");
76        }
77        println!("]");
78    }
79
80    Ok(())
81}
Source

pub fn compute_gradients( &self, data: &Array2<f64>, ) -> Result<(Array2<f64>, Array1<f64>)>

Compute gradients using contrastive divergence

Source

pub fn sample_hidden_given_visible( &self, visible: &ArrayView1<'_, f64>, ) -> Result<Array1<f64>>

Sample hidden units given visible units

Examples found in repository?
examples/quantum_boltzmann.rs (line 369)
364fn complete_pattern(rbm: &QuantumRBM, partial: &Array1<f64>) -> Result<Array1<f64>> {
365    // Use Gibbs sampling to complete pattern
366    let mut current = partial.clone();
367
368    for _ in 0..10 {
369        let hidden = rbm.qbm().sample_hidden_given_visible(&current.view())?;
370        current = rbm.qbm().sample_visible_given_hidden(&hidden)?;
371    }
372
373    Ok(current)
374}
Source

pub fn train( &mut self, data: &Array2<f64>, epochs: usize, batch_size: usize, ) -> Result<Vec<f64>>

Train the Boltzmann machine

Examples found in repository?
examples/quantum_boltzmann.rs (line 63)
47fn basic_qbm_demo() -> Result<()> {
48    // Create a small QBM
49    let mut qbm = QuantumBoltzmannMachine::new(
50        4,    // visible units
51        2,    // hidden units
52        1.0,  // temperature
53        0.01, // learning rate
54    )?;
55
56    println!("   Created QBM with 4 visible and 2 hidden units");
57
58    // Generate synthetic binary data
59    let data = generate_binary_patterns(100, 4);
60
61    // Train the QBM
62    println!("   Training on binary patterns...");
63    let losses = qbm.train(&data, 50, 10)?;
64
65    println!("   Training complete:");
66    println!("   - Initial loss: {:.4}", losses[0]);
67    println!("   - Final loss: {:.4}", losses.last().unwrap());
68
69    // Sample from trained model
70    let samples = qbm.sample(5)?;
71    println!("\n   Generated samples:");
72    for (i, sample) in samples.outer_iter().enumerate() {
73        print!("   Sample {}: [", i + 1);
74        for val in sample {
75            print!("{val:.0} ");
76        }
77        println!("]");
78    }
79
80    Ok(())
81}
Source

pub fn reconstruct(&self, visible: &Array2<f64>) -> Result<Array2<f64>>

Reconstruct visible units

Examples found in repository?
examples/quantum_boltzmann.rs (line 117)
84fn rbm_demo() -> Result<()> {
85    // Create RBM with annealing
86    let annealing = AnnealingSchedule::new(2.0, 0.5, 100);
87
88    let mut rbm = QuantumRBM::new(
89        6,    // visible units
90        3,    // hidden units
91        2.0,  // initial temperature
92        0.01, // learning rate
93    )?
94    .with_annealing(annealing);
95
96    println!("   Created Quantum RBM with annealing schedule");
97
98    // Generate correlated binary data
99    let data = generate_correlated_data(200, 6);
100
101    // Train with PCD
102    println!("   Training with Persistent Contrastive Divergence...");
103    let losses = rbm.train_pcd(
104        &data, 100, // epochs
105        20,  // batch size
106        50,  // persistent chains
107    )?;
108
109    // Analyze training
110    let improvement = (losses[0] - losses.last().unwrap()) / losses[0] * 100.0;
111    println!("   Training statistics:");
112    println!("   - Loss reduction: {improvement:.1}%");
113    println!("   - Final temperature: 0.5");
114
115    // Test reconstruction
116    let test_data = data.slice(s![0..5, ..]).to_owned();
117    let reconstructed = rbm.qbm().reconstruct(&test_data)?;
118
119    println!("\n   Reconstruction quality:");
120    for i in 0..3 {
121        print!("   Original:      [");
122        for val in test_data.row(i) {
123            print!("{val:.0} ");
124        }
125        print!("]  →  Reconstructed: [");
126        for val in reconstructed.row(i) {
127            print!("{val:.0} ");
128        }
129        println!("]");
130    }
131
132    Ok(())
133}
Source

pub fn sample_visible_given_hidden( &self, hidden: &Array1<f64>, ) -> Result<Array1<f64>>

Sample visible units given hidden units

Examples found in repository?
examples/quantum_boltzmann.rs (line 370)
364fn complete_pattern(rbm: &QuantumRBM, partial: &Array1<f64>) -> Result<Array1<f64>> {
365    // Use Gibbs sampling to complete pattern
366    let mut current = partial.clone();
367
368    for _ in 0..10 {
369        let hidden = rbm.qbm().sample_hidden_given_visible(&current.view())?;
370        current = rbm.qbm().sample_visible_given_hidden(&hidden)?;
371    }
372
373    Ok(current)
374}
Source

pub fn temperature(&self) -> f64

Get temperature

Examples found in repository?
examples/quantum_boltzmann.rs (line 188)
166fn energy_landscape_demo() -> Result<()> {
167    // Create small QBM for visualization
168    let qbm = QuantumBoltzmannMachine::new(
169        2,    // visible units (for 2D visualization)
170        1,    // hidden unit
171        0.5,  // temperature
172        0.01, // learning rate
173    )?;
174
175    println!("   Analyzing energy landscape of 2-unit system");
176
177    // Compute energy for all 4 possible states
178    let states = [
179        Array1::from_vec(vec![0.0, 0.0]),
180        Array1::from_vec(vec![0.0, 1.0]),
181        Array1::from_vec(vec![1.0, 0.0]),
182        Array1::from_vec(vec![1.0, 1.0]),
183    ];
184
185    println!("\n   State energies:");
186    for (i, state) in states.iter().enumerate() {
187        let energy = qbm.energy(state);
188        let prob = (-energy / qbm.temperature()).exp();
189        println!(
190            "   State [{:.0}, {:.0}]: E = {:.3}, P ∝ {:.3}",
191            state[0], state[1], energy, prob
192        );
193    }
194
195    // Show coupling matrix
196    println!("\n   Coupling matrix:");
197    for i in 0..3 {
198        print!("   [");
199        for j in 0..3 {
200            print!("{:6.3} ", qbm.couplings()[[i, j]]);
201        }
202        println!("]");
203    }
204
205    Ok(())
206}
Source

pub fn couplings(&self) -> &Array2<f64>

Get couplings matrix

Examples found in repository?
examples/quantum_boltzmann.rs (line 200)
166fn energy_landscape_demo() -> Result<()> {
167    // Create small QBM for visualization
168    let qbm = QuantumBoltzmannMachine::new(
169        2,    // visible units (for 2D visualization)
170        1,    // hidden unit
171        0.5,  // temperature
172        0.01, // learning rate
173    )?;
174
175    println!("   Analyzing energy landscape of 2-unit system");
176
177    // Compute energy for all 4 possible states
178    let states = [
179        Array1::from_vec(vec![0.0, 0.0]),
180        Array1::from_vec(vec![0.0, 1.0]),
181        Array1::from_vec(vec![1.0, 0.0]),
182        Array1::from_vec(vec![1.0, 1.0]),
183    ];
184
185    println!("\n   State energies:");
186    for (i, state) in states.iter().enumerate() {
187        let energy = qbm.energy(state);
188        let prob = (-energy / qbm.temperature()).exp();
189        println!(
190            "   State [{:.0}, {:.0}]: E = {:.3}, P ∝ {:.3}",
191            state[0], state[1], energy, prob
192        );
193    }
194
195    // Show coupling matrix
196    println!("\n   Coupling matrix:");
197    for i in 0..3 {
198        print!("   [");
199        for j in 0..3 {
200            print!("{:6.3} ", qbm.couplings()[[i, j]]);
201        }
202        println!("]");
203    }
204
205    Ok(())
206}

Auto Trait Implementations§

Blanket Implementations§

Source§

impl<T> Any for T
where T: 'static + ?Sized,

Source§

fn type_id(&self) -> TypeId

Gets the TypeId of self. Read more
Source§

impl<T> Borrow<T> for T
where T: ?Sized,

Source§

fn borrow(&self) -> &T

Immutably borrows from an owned value. Read more
Source§

impl<T> BorrowMut<T> for T
where T: ?Sized,

Source§

fn borrow_mut(&mut self) -> &mut T

Mutably borrows from an owned value. Read more
Source§

impl<T> From<T> for T

Source§

fn from(t: T) -> T

Returns the argument unchanged.

Source§

impl<T, U> Into<U> for T
where U: From<T>,

Source§

fn into(self) -> U

Calls U::from(self).

That is, this conversion is whatever the implementation of From<T> for U chooses to do.

Source§

impl<T> IntoEither for T

Source§

fn into_either(self, into_left: bool) -> Either<Self, Self>

Converts self into a Left variant of Either<Self, Self> if into_left is true. Converts self into a Right variant of Either<Self, Self> otherwise. Read more
Source§

fn into_either_with<F>(self, into_left: F) -> Either<Self, Self>
where F: FnOnce(&Self) -> bool,

Converts self into a Left variant of Either<Self, Self> if into_left(&self) returns true. Converts self into a Right variant of Either<Self, Self> otherwise. Read more
Source§

impl<T> Pointable for T

Source§

const ALIGN: usize

The alignment of pointer.
Source§

type Init = T

The type for initializers.
Source§

unsafe fn init(init: <T as Pointable>::Init) -> usize

Initializes a with the given initializer. Read more
Source§

unsafe fn deref<'a>(ptr: usize) -> &'a T

Dereferences the given pointer. Read more
Source§

unsafe fn deref_mut<'a>(ptr: usize) -> &'a mut T

Mutably dereferences the given pointer. Read more
Source§

unsafe fn drop(ptr: usize)

Drops the object pointed to by the given pointer. Read more
Source§

impl<T> Same for T

Source§

type Output = T

Should always be Self
Source§

impl<SS, SP> SupersetOf<SS> for SP
where SS: SubsetOf<SP>,

Source§

fn to_subset(&self) -> Option<SS>

The inverse inclusion map: attempts to construct self from the equivalent element of its superset. Read more
Source§

fn is_in_subset(&self) -> bool

Checks if self is actually part of its subset T (and can be converted to it).
Source§

fn to_subset_unchecked(&self) -> SS

Use with care! Same as self.to_subset but without any property checks. Always succeeds.
Source§

fn from_subset(element: &SS) -> SP

The inclusion map: converts self to the equivalent element of its superset.
Source§

impl<T, U> TryFrom<U> for T
where U: Into<T>,

Source§

type Error = Infallible

The type returned in the event of a conversion error.
Source§

fn try_from(value: U) -> Result<T, <T as TryFrom<U>>::Error>

Performs the conversion.
Source§

impl<T, U> TryInto<U> for T
where U: TryFrom<T>,

Source§

type Error = <U as TryFrom<T>>::Error

The type returned in the event of a conversion error.
Source§

fn try_into(self) -> Result<U, <U as TryFrom<T>>::Error>

Performs the conversion.
Source§

impl<V, T> VZip<V> for T
where V: MultiLane<T>,

Source§

fn vzip(self) -> V