/
AdaptiveProbMap.go
170 lines (134 loc) · 4.95 KB
/
AdaptiveProbMap.go
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
/*
Copyright 2011-2024 Frederic Langlet
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
you may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
*/
package entropy
//nolint (remove unused warnings)
import (
"errors"
kanzi "github.com/flanglet/kanzi-go/v2"
)
// AdaptiveProbMap maps a probability and a context to a new probability
// that the next bit will be 1. After each guess, it updates
// its state to improve future predictions.
type adaptiveProbMapData struct {
index int // last prob, context
rate uint // update rate
data []uint16 // prob, context -> prob
}
const (
LINEAR_APM = 0
LOGISTIC_APM = 1
FAST_LOGISTIC_APM = 2
)
type AdaptiveProbMap interface {
Get(bit int, pr int, ctx int) int
}
// LinearAdaptiveProbMap maps a probability and a context into a new probability
// using linear interpolation of probabilities
type LinearAdaptiveProbMap adaptiveProbMapData
// LogisticAdaptiveProbMap maps a probability and a context into a new probability
// using interpolation in the logistic domain
type LogisticAdaptiveProbMap adaptiveProbMapData
// FastLogisticAdaptiveProbMap is similar to LogisticAdaptiveProbMap but works
// faster at the expense of some accuracy
type FastLogisticAdaptiveProbMap adaptiveProbMapData
// NewBinaryEntropyEncoder creates an instance of AdaptiveProbMap
// given the provided type of APM.
func NewAdaptiveProbMap(mapType int, n, rate uint) (AdaptiveProbMap, error) {
if mapType == LINEAR_APM {
return newLinearAdaptiveProbMap(n, rate)
}
if mapType == LOGISTIC_APM {
return newLogisticAdaptiveProbMap(n, rate)
}
if mapType == FAST_LOGISTIC_APM {
return newFastLogisticAdaptiveProbMap(n, rate)
}
return nil, errors.New("Unknow APM type")
}
func newLogisticAdaptiveProbMap(n, rate uint) (*LogisticAdaptiveProbMap, error) {
this := &LogisticAdaptiveProbMap{}
size := n * 33
if size == 0 {
size = 33
}
this.data = make([]uint16, size)
this.rate = rate
for j := 0; j <= 32; j++ {
this.data[j] = uint16(kanzi.Squash((j-16)<<7) << 4)
}
for i := uint(1); i < n; i++ {
copy(this.data[i*33:], this.data[0:33])
}
return this, nil
}
// get returns improved prediction given current bit, prediction and context
func (this *LogisticAdaptiveProbMap) Get(bit int, pr int, ctx int) int {
// Update probability based on error and learning rate
g := (-bit & 65528) + (bit << this.rate)
this.data[this.index+1] += uint16((g - int(this.data[this.index+1])) >> this.rate)
this.data[this.index] += uint16((g - int(this.data[this.index])) >> this.rate)
pr = kanzi.STRETCH[pr]
// Find index: 33*ctx + quantized prediction in [0..32]
this.index = ((pr + 2048) >> 7) + 33*ctx
// Return interpolated probability
w := pr & 127
return (int(this.data[this.index+1])*w + int(this.data[this.index])*(128-w)) >> 11
}
func newFastLogisticAdaptiveProbMap(n, rate uint) (*FastLogisticAdaptiveProbMap, error) {
this := &FastLogisticAdaptiveProbMap{}
this.data = make([]uint16, n*32)
this.rate = rate
for j := 0; j < 32; j++ {
this.data[j] = uint16(kanzi.Squash((j-16)<<7) << 4)
}
for i := uint(1); i < n; i++ {
copy(this.data[i*32:], this.data[0:32])
}
return this, nil
}
// get returns improved prediction given current bit, prediction and context
func (this *FastLogisticAdaptiveProbMap) Get(bit int, pr int, ctx int) int {
// Update probability based on error and learning rate
g := (-bit & 65528) + (bit << this.rate)
this.data[this.index] += uint16((g - int(this.data[this.index])) >> this.rate)
this.index = ((kanzi.STRETCH[pr] + 2048) >> 7) + 32*ctx
return int(this.data[this.index]) >> 4
}
func newLinearAdaptiveProbMap(n, rate uint) (*LinearAdaptiveProbMap, error) {
this := &LinearAdaptiveProbMap{}
size := n * 65
if size == 0 {
size = 65
}
this.data = make([]uint16, size)
this.rate = rate
for j := 0; j <= 64; j++ {
this.data[j] = uint16(j<<6) << 4
}
for i := uint(1); i < n; i++ {
copy(this.data[i*65:], this.data[0:65])
}
return this, nil
}
// get returns improved prediction given current bit, prediction and context
func (this *LinearAdaptiveProbMap) Get(bit int, pr int, ctx int) int {
// Update probability based on error and learning rate
g := (-bit & 65528) + (bit << this.rate)
this.data[this.index+1] += uint16((g - int(this.data[this.index+1])) >> this.rate)
this.data[this.index] += uint16((g - int(this.data[this.index])) >> this.rate)
// Find index: 65*ctx + quantized prediction in [0..64]
this.index = (pr >> 6) + 65*ctx
// Return interpolated probability
w := pr & 127
return (int(this.data[this.index+1])*w + int(this.data[this.index])*(128-w)) >> 11
}